ADVERTIMENT. Lʼaccés als continguts dʼaquesta tesi queda condicionat a lʼacceptació de les condicions dʼús establertes per la següent llicència Creative Commons: http://cat.creativecommons.org/?page_id=184

ADVERTENCIA. El acceso a los contenidos de esta tesis queda condicionado a la aceptación de las condiciones de uso establecidas por la siguiente licencia Creative Commons: http://es.creativecommons.org/blog/licencias/

WARNING. The access to the contents of this doctoral thesis it is limited to the acceptance of the use conditions set by the following Creative Commons license: https://creativecommons.org/licenses/?lang=en Doctorand: Mauricio Tobón Restrepo Directores: Yvonne Espada Gerlach & Rosa Novellas Torroja

Tesi Doctoral

Barcelona, 29 de juliol de 2016

This thesis has received financial support from the Colombian government through the “Francisco José de Caldas” scholarship program of COLCIENCIAS and from the Corporación Universitaria Lasallista.

DEDICATED TO

A los que son la razón y la misión de esta tesis… LOS GATOS.

A mis padres y hermanos.

A Ismael.

Vor mijn poffertje.

ACKNOWLEDGMENTS

Tal vez es la parte que se pensaría más fácil de escribir, pero sin duda se juntan muchos sentimientos al momento de mirar atrás y ver todo lo que has aprendido y todas las personas que han estado a tu lado dándote una palabra de aliento… y es ahí cuando se asoma la lágrima…

Sin duda alguna, comienzo agradeciendo a los propietarios de todos los gatos incluidos en este estudio, sin ellos esto no habría sido posible.

A continuación agradezco a mis directoras de tesis, la Dra. Rosa Novellas y la Dra. Yvonne Espada. Muchas gracias por creer en mí, por apoyarme y por tenerme tanta paciencia. He aprendido mucho de vosotras y por siempre les estaré muy agradecido.

A mis patrocinadores en Colombia: Colciencias y La Corporación Universitaria Lasallista. Sin su apoyo económico, este sueño no se habría cumplido.

Al servei de diagnòstic per l’imatge del Hospital Clínic Veterinari de la Universitat Autònoma de Barcelona. Judith Saura, Raúl Altuzarra, Carlo Anselmi, Eli Domínguez. Muchas gracias por vuestra amistad, compañerismo, enseñanzas, apoyo y confianza. Toda vuestra ayuda en gran parte han hecho posible esta tesis. También os agradezco el que me hayáis contagiado del mundo del diagnóstico por imagen que se ha convertido en mi sueño y mi pasión. Os quiero montones.

To Maxim, your love and support, your patience helped me to keep this through and get the strength to finish it. Dank je wel schatje.

A todo el personal del Hospital Clínic Veterinari de la Universitat Autònoma de Barcelona por el apoyo logístico, las palabras de ánimo, la ayuda en conseguir los pacientes y el cuidado ofrecido a los mismos durante su recuperación de la anestesia.

Al servicio de anestesia del hospital, especialmente a Adrià Aguilar y Xavier Moll; siempre dispuestos para ayudarme de muy buena gana. Han sido muchas horas y muchos momentos de incertidumbre que hemos pasado juntos evaluando a todos los gatos, aprecio mucho todo lo que me apoyaron y aguantaron.

A los servicios de bioquímica y hematología de la facultad de veterinaria por vuestra diligencia al procesar las muestras y en la lectura de las citologías. Montse, Toni y la Dra. Cuenca.

A todos mis amigos en Colombia, Henry Daniel Espinosa, Juan Fernando Suárez, Héctor Buitrago, Luis Andrés Vergas, mis colegas veterinarios y mis compañeros de trabajo, por todo el apoyo y la compañía en la distancia.

A mis amigos en España, Miquel Ciurana, Miquel Falla, Juan Luis Castillo, Robert Kempton, José Rodríguez, Ignasi Oliver, Fernando y Joan Franch vuestro apoyo y amistad es una de las mejores cosas que me ha pasado en mi estancia en Barcelona. Gracias por enseñarme vuestra cultura, por abrirme vuestros hogares y ayudarme a reconocer que en este mundo hay personas como vosotros que desinteresadamente se ayudan las unas a las otras. Os quiero montones.

A Alejandro Durá, la amistad verdadera y el cariño sincero nacen de un corazón tan puro como el tuyo. Me has apoyado mucho más de lo que te puedas imaginar, siempre con tus palabras de ánimo, con tus consejos y tus soluciones a lo imprevisto. Estoy seguro que puedo seguir contando contigo para siempre. Te quiero un montón.

Y por último, y no porque sea menos importante, sino por el contrario porque son mi vida entera, a mis padres y hermanos, a mi precioso sobrino Ismael. Esta tesis es por ustedes y el estar lejos durante tantos años ha sido un alto precio a pagar. Los amo demasiado y los extraño cada segundo que paso lejos de casa.

Index ABBREVIATIONS ...... 19 1. INTRODUCTION ...... 23 2. LITERATURE REVIEW ...... 27 2.1. Anatomy of the in the domestic cat ...... 29 2.1.1. Lymph nodes (lymphonodi) ...... 29 2.1.2. Lymph centers (lymphocentra) ...... 30 a. Head and neck lymph centers ...... 30 • Parotid lymph center (lymphocentrum parotideum) ...... 31 • Mandibular lymph center (lymphocentrum mandibulare) ...... 31 • Retropharyngeal lymph center (lymphocentrum retropharyngeum) ...... 31 • Superficial cervical lymph center (lymphocentrum cervicale superficiale) .... 32 • Deep cervical lymph center (lymphocentrum cervicale profundum) ...... 32 b. Forelimb lymph center ...... 33 • Axillary lymph center (lymphocentrum axillare) ...... 33 c. Thoracic lymph centers ...... 34 • Dorsal thoracic lymph center (lymphocentrum thoracicum dorsale) ...... 34 • Ventral thoracic lymph center (lymphocentrum thoracicum ventrale) ...... 35 • Mediastinal lymph center (lymphocentrum mediastinale) ...... 36 • Bronchial lymph center (lymphocentrum bronchale) ...... 36 d. Abdominal and pelvic lymph centers ...... 37 • Celiac lymph center (lymphocentrum celiacum) ...... 37 • Cranial mesenteric lymph center (lymphocentrum mesentericum craniale) 38 • Caudal mesenteric lymph center (lymphocentrum mesentericum caudale) . 39 • Lumbar lymph center (lymphocentrum lumbale) ...... 39 • Iliosacral lymph center (lymphocentrum iliosacrale) ...... 41 • Inguinofemoral lymph center (lymphocentrum inguinofemorale) ...... 42 • Ischiatic lymph center (lymphcentrum ischiadicum) ...... 43 e. Hindlimb lymph centers ...... 43 • Iliofemoral lymph center (lymphocentrum iliofemorale) ...... 44 • Popliteal lymph center (lymphocentrum popliteum) ...... 44 2.2. Ultrasonography of the lymphatic system ...... 44 2.2.1. Physical principles ...... 45

a. Propagation of sound ...... 45 b. Frequency ...... 46 c. Wavelength ...... 46 d. Reflection ...... 46 e. Refraction ...... 46 f. Acoustic impedance ...... 47 2.2.2. Modes of image display ...... 47 a. B-mode (Brightness mode, B-scan or gray scale) ...... 47 2.2.3. Instruments ...... 48 2.2.4. Ultrasonography of the normal lymph nodes ...... 48 2.2.5. Ultrasonographic features of abnormal lymph nodes ...... 50 2.3. Computed tomography of the lymphatic system ...... 51 2.3.1. CT principles ...... 51 2.3.2. CT scans components ...... 52 a. Gantry ...... 52 b. X-ray tube ...... 52 c. Collimator ...... 52 d. Detector system ...... 53 e. Patient table ...... 53 f. Work station ...... 53 2.3.3. Image acquisition principles ...... 54 a. Scan field of view (SFOV) ...... 54 b. Display field of view (DFOV) ...... 54 c. Slice thickness ...... 54 d. Slice interval ...... 55 e. Helical pitch ...... 55 2.3.4. Image reconstruction ...... 55 a. Voxel (Volume element) ...... 55 b. Pixel (Picture element) ...... 56 c. Hounsfield Units (HU) ...... 56 d. Window width ...... 56 e. Window level ...... 56 2.3.5. Multiplanar reconstruction of images ...... 57 2.3.6. Contrast studies ...... 57 2.3.7. Computed tomography of normal lymph nodes ...... 58 2.3.8. Computed tomography of abnormal lymph nodes in the cat ...... 59 3. OBJECTIVES ...... 61 4. STUDIES ...... 65 4.1. Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, , and forelimb...... 67 4.2. Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb...... 105 4.3. Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography...... 153 5. GENERAL DISCUSSION ...... 185 6. CONCLUSIONS ...... 193 7. SUMMARY ...... 197 8. RESUM ...... 201 9. RESUMEN ...... 205 10. OTHER STUDIES DERIVED FROM THIS THESIS ...... 209 11. REFERENCES ...... 247

ABBREVIATIONS

ALT Alanine-aminotransferase AAxLN Accessory axillary AxLN Axillary lymph node CDCLN Caudal deep cervical lymph node CELN Caudal epigastric lymph node CoLN Colic lymph node CrMLN Cranial mediastinal lymph node CMLN Caudal mesenteric lymph node CT Computed tomography DDi-UU Imaging Division of the veterinary faculty of Utrecht University DFOV Display field of view DSCLN Dorsal superficial cervical lymph node FeLV Feline leukemia virus FHCV Fundació Hospital Clínic Veterinari FIP Feline infectious peritonitis FIV Feline immunodeficiency virus FNA Fine needle aspiration GGT Gamma-glutamyl-transferase GLN Gastric lymph node HLN Hepatic lymph node HU Hounsfield units ICLN Ileocecal lymph node InILN Internal iliac lymph node IsLN Ischiatic lymph node JLN Jejunal lymph node LALN Lumbar aortic lymph node LC Lymph center LL Left lateral LM Left medial LMR Left medial retropharyngeal LMRLN Left medial retropharyngeal lymph node LN Lymph node LTBLN Left tracheobronchial lymph node

MHz Megahertz MILN Medial iliac lymph node MnLN MPR Multiplanar reconstruction MRLN Medial retropharyngeal lymph node MRI Magnetic resonance imaging MTBLN Middle tracheobronchial lymph node NAV Nomina Anatomica Veterinaria PCR Polymerase chain reaction PET Positron emission tomography PLN Parotid lymph node PdLN Pancreaticoduodenal lymph node PoLN Popliteal lymph node RL Right lateral RM Right medial RMR Right medial retropharyngeal RMRLN Right medial retropharyngeal lymph node ROI Region of interest RTBLN Right tracheobronchial lymph node SaLN Sacral lymph node SCLN Superficial cervical lymph node SD Standard deviation SFOV Scan field of view SILN Superficial inguinal lymph node SLN Sternal lymph node SpLN Splenic lymph node UAB Universitat Autònoma de Barcelona US Ultrasonography VSCLN Ventral superficial cervical lymph node

1. INTRODUCTION

INTRODUCTION 25

1. INTRODUCTION

As part of the lymphatic and circulatory system, the lymph nodes play an important role in the diagnosis of neoplastic and infectious diseases.

In human medicine, lymph nodes are important structures that play a major role in tumor staging, choice of the therapy and predicting outcome of malignant diseases. In the last decades, invasive methods like surgical dissection of lymph nodes, serial histologic examination (fine needle aspirates) and even adjuvant chemotherapy have been considered the gold standard and efficient diagnostic tools. However, these techniques are expensive, time consuming and carry a significant risk of complications such as seromas, lymphedema, inflammation, and cosmetic alterations. Due to these facts, researchers have been focused (especially for neoplastic disease) in the improvement and understanding of non-invasive imaging techniques. So far, computed tomography (CT), magnetic resonance imaging (MRI) using lymphotropic contrast agents (USPIOs), and contrast enhanced ultrasonography (US) allowed a fair noninvasive assessment of lymph nodes in humans (Wunderbaldinger, 2006).

Previous studies in veterinary medicine have shown that US and CT are useful techniques to assess the lymph nodes, especially in dogs. In 2004, Llabres-Diaz characterized the medial iliac lymph nodes in dogs and provided ultrasonographic features to be used as normal reference. The lack of available information regarding these lymph nodes in cats has caused the canine features to be applied to this species. Some information regarding imaging features of the feline lymph nodes are available. However, even though literature characterizing some of these feline lymph nodes is available, normal values for size and ability to depict many of them, especially with advance imaging techniques, such as CT, are lacking. This makes interpretation of studies in clinical patients difficult. Normal feline abdominal lymph nodes have been assessed with US (Schreurs et al., 2008). Normal retropharyngeal lymph nodes have been assessed using US and CT

26 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

(Nemanic & Nelson, 2012). Sentinel lymph nodes of normal mammary glands have been assessed using radiography and CT indirect lymphography (Patsikas et al., 2010). Nevertheless, there are some limitations in these studies, including small sample size (abdominal study made by Schreurs et al., 2008), not using contrast medium (retropharyngeal study (Nemanic & Nelson, 2012)), and assessment of only some of the body lymph centers (mammary lymph nodes study) (Patsikas et al., 2010). Therefore further studies involving the assessment of all the body lymph centers would be necessary to further characterize them.

To the authors knowledge there are no complete reports about normal features and assessment of all the feline lymph nodes using US and CT. On the other hand, only a limited number of studies comparing neoplastic and infectious diseases with normal lymph nodes are available, and there are no studies comparing the sensibility and specificity of US versus CT to differentiate normal from abnormal lymph nodes in cats.

2. LITERATURE REVIEW

LITERATURE REVIEW 29

2. LITERATURE REVIEW

2.1. Anatomy of the lymphatic system in the domestic cat

The study of the lymphatic system anatomy has leaded to improve our knowledge of its structure and function. According to Tompkins (1993), the lymphatic system is divided into primary and secondary lymphoid organs. The thymus and the bone marrow are the primary components. The , lymph nodes, aggregated lymphoid tissue (tonsils), and lymphatic vessels are the secondary components. Another classification of the lymphatic system has been made by Bezuidenhout (2013) dividing it into a cellular and a vascular component. All the organs with lymphatic tissue and lymph nodes are the cellular component. The lymph capillaries, vessels, and ducts are the vascular component. The lymphatic system‘s main functions are the immunological defense of the body and the draining and filtration of the lymph (Bezuidenhout, 2013; Dyce, Sack, & Wensing, 2002; Tompkins, 1993).

2.1.1. Lymph nodes (lymphonodi)

A lymph node (LN) is a small solid mass of lymphatic tissue (NAV, 2012). Normally, lymph nodes (LNs) are round or bean shaped structures located in the course of the lymphatic vessels (Tompkins, 1993). The LNs are considered to be the structural and functional unit of the lymphatic system (Bezuidenhout, 2013). They have a convex surface, which receives lymphatic afferent vessels, and a concave surface called the hilum. Bezuidenhout, (2013) described the hilum as the region of the lymph node that receives the blood supply and where the lymphatic efferent vessels leave the lymph node.

30 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

The LNs are named by regional location, and the number of nodes in a region varies among the species; in cats, the number, size, and presence of some LNs shows individual variations (Tompkins, 1993).

The LNs have two important functions. A lymph filter role that traps material received from the lymphatic afferent vessels, and a lymphocytes germinal center, providing a site for the immune response to develop (Bezuidenhout, 2013; Tompkins, 1993). The LNs have a special distribution in the body. Normally, they are located in places that provide maximum protection but without interfering in the functioning of the skeletal, muscular, and circulatory systems. It is common to find them in the adipose tissue, in the flexor angles of the joints, in the mediastinum, in the mesentery, and in the angles formed by the origin of many of the large blood vessels (Bezuidenhout, 2013).

The normal histology of a LN shows a capsule containing elastic and smooth muscle fibers, with an internal framework consisting of septa and trabeculae that sustain several lymph nodules. These lymph nodules are the structural unit of the lymph node, and contain the germinal center. They are mainly distributed in the cortex (Bezuidenhout, 2013).

2.1.2. Lymph centers (lymphocentra)

The Nomina Anatomica Veterinaria defines a lymph center as a lymph node or a group of lymph nodes that occurs in the same region of the body and receives afferent vessels from approximately the same region in most domestic mammals (NAV, 2012, p. 116).

a. Head and neck lymph centers

Three lymph centers have been reported in the head of domestic mammals (Figure 1): parotid, mandibular, and retropharyngeal; but only two have been

LITERATURE REVIEW 31

described for the neck: superficial cervical and deep cervical. These are also present in the cat.

• Parotid lymph center (lymphocentrum parotideum)

It is formed by 1 to 2 parotid LNs located at the cranial border of the parotid salivary gland, in contact with the superficial temporal vein. The afferent lymphatic vessels come from the dorsal aspect of the ear, the head, and neck; also lymph from the eyelids, the lips, and the parotid salivary gland drains to the parotid LN. The efferent drainage goes to the lateral retropharyngeal LN (NAV, 2012; Saar & Getty, 1982; Tompkins, 1993).

• Mandibular lymph center (lymphocentrum mandibulare)

Tompkins (1993) reported 2 mandibular LNs that form this lymph center; however, the NAV (2012) also report the presence of a mandibular accessory LN for this lymph center in the cat. The mandibular LNs lay medial and lateral to the facial vein at the angle of the mandibular bone, deep to the platysma muscle. They drain lymph from the lips, chin, lower jaw, eyes, part of the oral cavity, buccal salivary glands, and skin from the mandibular region. The efferent lymphatic vessels go to the medial retropharyngeal LN (Saar & Getty, 1982; Tompkins, 1993).

• Retropharyngeal lymph center (lymphocentrum retropharyngeum)

This consists of medial and lateral retropharyngeal LNs. Only one medial retropharyngeal lymph node has been reported and it is located adjacent to the carotid sheath at the level of the atlas and it is bounded medially by the pharyngeal musculature. It receives the lymph from the tongue, oral and nasal passages, thyroids, mandibular salivary glands, mandibular LNs, parotid LNs, and lateral retropharyngeal LNs. Its efferent vessels form the jugular trunk. A number of 3 to 4 lateral retropharyngeal LNs have been reported in most cats. They are located caudal to the parotid salivary gland

32 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

along the caudal auricular vein. They receive lymph that comes from the ear and parotid gland and drain to the medial retropharyngeal lymph node (NAV, 2012; Tompkins, 1993).

• Superficial cervical lymph center (lymphocentrum cervicale superficiale)

There are one ventral and two dorsal superficial cervical LNs in the cat. Tompkins (1993) described the location of the dorsal LNs deep to the omotransversarius muscle and cranioventrally to the trapezius muscle. They receive lymph from the dorsal part of the neck and from the forelimbs. Their efferent vessels go into the jugular trunk, tracheal trunk, and into the left side of the thoracic duct (Saar & Getty, 1982). The ventral superficial cervical LN lies on the external jugular vein, near the junction of the superficial cervical vein. It drains lymph from the ventral part of the neck and the thoracic inlet and its efferent vessels reach the jugular trunk, the tracheal trunk, or the thoracic duct (Saar & Getty, 1982; Tompkins, 1993).

• Deep cervical lymph center (lymphocentrum cervicale profundum)

It consists of the middle deep cervical (inconsistently present in cats) and the caudal deep cervical LNs located at the ventral surface of the trachea near the thoracic inlet. Their drainage areas are the thyroid glands, trachea, and cervical . Their efferent vessels go into the tracheal trunk (Saar & Getty, 1982; Tompkins, 1993). Bezuidenhout (2013) reported the presence of cranial deep in 30% of dogs. However, this LN has been described only in 1% of the cats, located in the ventral aspect of the 20th to 23rd tracheal cartilages along the internal jugular vein (Sugimura, Kudo, & Takahata, 1955, 1959).

LITERATURE REVIEW 33

Figure 1. Superficial lymph centers in the cat. 1. Parotid. 2a & b. Mandibular (a. medial; b. lateral). 3 a & b. Retropharyngeal (a. medial; b. lateral). 4. Deep cervical. 5 a & b. Superficial cervical (a. dorsal; b. ventral). 6 a & b. Axillary (a. axillary LN; b. accessory axillary LN). 7a & b. Inguinofemoral (a. caudal epigastric LN; b. superficial inguinal LN). 8. Ischiatic. 9. Popliteal. Adapted from a free picture in the website: https://nl.pinterest.com/pin/164240717636259107/

b. Forelimb lymph center

There is only one lymph center that drains all the lymph from the forelimb.

• Axillary lymph center (lymphocentrum axillare)

It consists of the axillary and accessory axillary LNs. In cats, Tompkins (1993) reports the presence of only one axillary LN on each side. It is located in the medial surface of the forelimb between the axillary and lateral thoracic veins. It drains the medial aspect of the forelimb and the lateral thoracic wall. Its efferent vessels end in the jugular venous angle. Three to 5 accessory axillary LNs are most commonly identified at the medial aspect of the latissimus dorsi muscle between the 3rd and 6th intercostal space along the lateral thoracic vein. The accessory axillary LNs drain the medial and cranial aspect of the forelimb, the skin of the lumbar region, the lateral chest walls, and the cranial half of the mammary glands. The efferent vessels end in the

34 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

axillary lymph node (Saar & Getty, 1982; Sugimura, Kudo, & Takahata, 1956).

c. Thoracic lymph centers

Four lymph centers have been described in the thorax (Figure 2). They can be divided into parietal and visceral lymph centers. The ventral and dorsal thoracic lymph centers are considered to be part of the parietal group. Mediastinal and bronchial lymph centers are considered to be part of the visceral group.

Figure 2. Thoracic lymph centers in the cat. 1. Cranial mediastinal. 2 a-d. Ventral thoracic (a. Cranial sternal LN; b. Caudal sternal LN; c. Superficial cranial epigastric LN; d. Phrenic LN). 3 a & b. Dorsal thoracic (a. Intercostal LNs; b. Aortic thoracic LNs). 4 a-c. Bronchial (a. Left/right tracheobronchial LNs; b. Middle tracheobronchial LN; c. Pulmonary LN).

• Dorsal thoracic lymph center (lymphocentrum thoracicum dorsale)

This lymph center is represented by two different groups of LNs: ∼ The aortic thoracic LNs: round in shape and located along the azygos vein and the Aorta on the right side, at the ventral aspect of the body of the thoracic vertebras (Saar & Getty, 1982; Sugimura et al., 1959). This group of 1 – 5 LNs, drains the costal pleura and peritoneum, and

LITERATURE REVIEW 35

their efferent vessels end in the thoracic duct (Saar & Getty, 1982; Tompkins, 1993). ∼ The intercostal LNs: this group of nodes can be found inconsistently at the level of the vertebral end of the intercostal spaces in contact with the intercostal vessels. Spheroidal in shape, they drain the lymph from the dorsal costal pleura. Some efferent vessels might go to the aortic thoracic LNs, or end directly in the thoracic duct (Saar & Getty, 1982; Sugimura et al., 1959).

• Ventral thoracic lymph center (lymphocentrum thoracicum ventrale)

This lymph center is represented by three groups of lymph nodes according to the NAV (2012): ∼ Sternal LNs: the cranial sternal LN is the lymph node of this group that is most frequently found in cats. It lies along the internal thoracic vessels at the level of the second costal cartilage. Small nodes may be found caudally until the sixth costal cartilage covered by the transversus thoracis muscle (Saar & Getty, 1982; Sugimura et al., 1959). A second LN has been described between the pericardial branches of the thoracic internal vein in the pericardial vertex and is known as caudal sternal LN (NAV, 2012; Saar & Getty, 1982). These lymph nodes drain the ventral costal pleura, diaphragm, pericardium, heart, the cranioventral portion of the abdominal wall, and the cranial mediastinal LNs. The efferent vessels go to the thoracic duct and jugular trunk (Saar & Getty, 1982). ∼ Superficial cranial epigastric LN (former xiphoid LN): In the cat, and also in the rabbit, this node has been described as being located deep to the rectus abdominis muscle, along the cranial epigastric vein and caudally to the xiphoid cartilage (NAV, 2012; Saar & Getty, 1982; Sugimura et al., 1956). It drains the cranial abdominal wall and the efferent vessels have not been described (Saar & Getty, 1982; Sugimura et al., 1956).

36 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

∼ Phrenic LN: lies in the diaphragmatic pleura near the foramen venae cavae. Spherical in shape, this LN drains the diaphragm and its efferent vessels end at the sternal LN or mediastinal or bronchial lymph centers (Saar & Getty, 1982; Sugimura et al., 1959).

• Mediastinal lymph center (lymphocentrum mediastinale)

It is only represented by one group of 2 to 8 small and ellipsoid lymph nodes reported as cranial mediastinal LNs, which are located along the cranial vena cava and ventral aspect of the trachea and esophagus. These nodes drain the heart, trachea, thymus, and other LNs from the thoracic cavity. The efferent vessels end at the thoracic duct and the jugular trunk (Saar & Getty, 1982; Tompkins, 1993).

• Bronchial lymph center (lymphocentrum bronchale)

This lymph center is represented by two groups of LNs, as follows: ∼ Tracheobronchial LNs: the position of this group is related to the tracheal bifurcation. In cats, at the cranial aspect of the main right bronchus lies the right tracheobronchial LN that drains lymph from the right lung, and in some cases from the pericardium, esophagus, and right pulmonary LN. Its efferent vessels end in the cranial mediastinal LNs. At the cranial aspect of the main left bronchus lies the left tracheobronchial LN that drains lymph from the left lung, and as in the right side, and less frequently, may receives lymphatic vessels from the pericardium, esophagus, and left pulmonary, middle tracheobronchial and phrenic LNs. Its efferent vessels end also in the cranial mediastinal LNs. At the caudal aspect of the carina lies the middle tracheobronchial LN. The afferent vessels come mainly from the lungs, diaphragm, and esophagus. Its efferent vessels end at the left tracheobronchial LN, cranial mediastinal LNs, lymphatic trunk, or directly at the cranial vena cava (Saar & Getty, 1982; Sugimura et al., 1959).

LITERATURE REVIEW 37

∼ Pulmonary LNs: as in dogs, this group of nodes is often absent in cats. Small spherical nodes were described by Sugimura et al. (1959) on the dorsal surface of the main bronchi, between the pulmonary tissue and the tracheobronchial LNs. These nodes drain the lungs and their efferent vessels go to the tracheobronchial LNs (Bezuidenhout, 2013; Saar & Getty, 1982; Sugimura et al., 1959).

d. Abdominal and pelvic lymph centers

The lymph centers in these regions, as in the thorax, can be divided into visceral and parietal groups. Getty et al. (1982) reported three lymph centers for the visceral group: celiac, cranial mesenteric, and caudal mesenteric (Figure 3); and four lymph centers for the parietal group: lumbar, iliosacral, inguinofemoral, and ischiatic (Figures 1 & 4).

• Celiac lymph center (lymphocentrum celiacum)

This lymph center is represented by four groups of LNs: ∼ Splenic LNs: Can be 1 – 3 in number and located along the splenic vein and its junction with the short gastric vein (Saar & Getty, 1982; Sugimura, Kudo, & Takahata, 1958). Spherical or ellipsoid in shape, these LNs drain the spleen, greater gastric curvature, including the cardia and fundus, and the (Getty et al., 1982; Sugimura et al., 1958; Tompkins, 1993). Afferent vessels may also come from gastric, pancreaticoduodenal, hepatic, and jejunal LNs (Saar & Getty, 1982; Sugimura et al., 1958). Efferent vessels drain into the celiac trunk. ∼ Gastric LNs: one to 4 in number, these LNs are located at each side of the lesser curvature of the . They drain the stomach and its efferent vessels go to the splenic and hepatic LNs (Saar & Getty, 1982; Sugimura et al., 1958; Tompkins, 1993). ∼ Hepatic LNs: the existence of 1 to 6 hepatic LNs have been reported (Getty et al.,1982; Sugimura et al., 1958; Tompkins, 1993). They are

38 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

located around the junction of the gastroduodenal vein with the portal vein. Their drainage area is the liver, the stomach, the cranial duodenum, gastric LNs, and pancreaticoduodenal LNs. The efferent vessels drain into the celiac trunk (Saar & Getty, 1982; Tompkins, 1993). ∼ Pancreaticoduodenal LN: just 1 in number, this node is located around the junction of the right gastroepiploic and cranial pancreaticoduodenal veins in the caudal part of the pylorus. The drainage area is the pancreas, cranial portion of the duodenum, and greater curvature of the stomach. The efferent vessels drain into the hepatic and, occasionally, into the (Saar & Getty, 1982; Tompkins, 1993).

• Cranial mesenteric lymph center (lymphocentrum mesentericum craniale)

Three groups of LNs represent this lymph center as follows: ∼ Jejunal LNs: these LNs are located in the mesojejunum along the jejunal branch of the cranial mesenteric artery. There are normally 5 – 6 nodes but the number may vary from 2 – 20 (Saar & Getty, 1982; Sugimura et al., 1958; Tompkins, 1993). They drain the jejunum, ileum, caudal portion of the duodenum, and pancreas. The efferent vessels contribute to form the intestinal trunk that joins the celiac trunk to form the visceral trunk that ends at the (Saar & Getty, 1982). ∼ Ileocecal LNs: situated along the cecal branches from the ileocolic vessels at both sides of the cecum. Normally there is one LN to each side but two have been also reported, and sometimes they can be absent. They drain the cecum and the last portion of the ileum. The efferent vessels drain into the right colic LNs (NAV, 2012; Saar & Getty, 1982; Sugimura et al., 1958; Tompkins, 1993). ∼ Colic LNs: according to the NAV, the term Lnn. colici refers to all lymph nodes that are located along the various segments of the colon, except those that lie directly at the ileocolic junction (NAV, 2012, p.

LITERATURE REVIEW 39

117). Getty et al. (1982) reported that those located at the ileocolic junction in relation to the branches of the ileocolic artery, deep in the mesocolon, are the right colic LNs. The number of nodes in this group varies frequently (from 1 to 14), but in general 4 to 5 nodes are normally seen (Saar & Getty, 1982; Tompkins, 1993). The afferent vessels come from the ascending and transverse colon, ileum, cecum, and ileocecal lymph nodes. The efferent vessels end into the intestinal trunk (Saar & Getty, 1982; Sugimura et al., 1958; Tompkins, 1993).

• Caudal mesenteric lymph center (lymphocentrum mesentericum caudale)

It is only represented by the caudal mesenteric LNs. Normally 2 to 3 nodes can be found located at the bifurcation of the caudal mesenteric artery into cranial and caudal branches. The afferent vessels come from the descending colon and rectus. The efferent vessels drain into the medial iliac LNs, lumbar aortic LNs, or lumbar trunk (Saar & Getty, 1982).

• Lumbar lymph center (lymphocentrum lumbale)

As in canines, this lymph center consists of the lumbar aortic and renal lymph nodes in the cat. ∼ Lumbar aortic LNs: they are located along the aorta and caudal vena cava, from the diaphragm to the deep circumflex iliac arteries. Sugimura et al. (1958) reported the existence of 4 – 11 nodes in 19/24 cats. According to Getty et al. (1982) and Tompkins (1993), afferent vessels drain the testes, ovaries (occasionally uterus and splenic, colic, and hepatic LNs), the dorsal wall of the abdomen, diaphragm, medial and internal iliac LNs, and the kidneys. The efferent vessels end into the lumbar trunk. ∼ Renal LNs: located in relationship with the renal vessels, these LNs are difficult to differentiate from the lumbar aortic LNs. Getty et al. (1982) reported the existence of 3 to 4 nodes. The afferent vessels drain the kidneys, adrenal glands, peritoneum, abdominal wall, and

40 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

diaphragm, and occasionally the celiac lymph center, ovaries, and testes. The efferent vessels end into the cisterna chyli.

Figure 3. Diagram of the abdominal LNs and their relative anatomical positions. 1. Hepatic; 2. Gastric; 3. Splenic; 4. Pancreaticoduodenal; 5. Iliocaecal; 6. Colic; 7. Jejunal; 8. Caudal mesenteric; L: Liver; G: Gallbladder; St: Stomach; D: Duodenum; Sp: Spleen; P: Pancreas; J: Jejunum; Ca: Cecum; Co: Colon, R: Rectum; LK: Left kidney. Adapted from the original drawing of José Rodriguez.

LITERATURE REVIEW 41

Figure 4. Lumbar & iliosacral lymph centers in the cat. 1. Renal. 2. Lumbar aortic. 2. Medial iliac. 3. Internal iliac 4. Sacral. RK: Right kidney. LK: Left kidney (Figure adapted from Diagnóstico ecográfico en el gato. Rosa Novellas et al. (Servet editorial, 2015))

• Iliosacral lymph center (lymphocentrum iliosacrale)

In cats, this lymph center consists of the medial iliac, internal iliac, and sacral LNs. ∼ Medial iliac LNs: two to five in number, these nodes are located bilaterally along the aorta between the deep circumflex iliac and external iliac arteries. They receive afferent lymph from the uterus,

42 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

medial and caudal portions of the abdominal wall, hindlimbs, pelvis, tail and iliofemoral, inguinofemoral and ischiatic lymph centers. The efferent vessels form the lumbar trunks, and end into the cisterna chyli (NAV, 2012; Saar & Getty, 1982; Sugimura et al., 1958; Tompkins, 1993). ∼ Internal iliac LNs: were formerly known as hypogastric LNs. They are located in relation to the internal iliac arteries and 2 – 3 nodes have been described. They drain lymph from the rectum, uterine body, bladder, the tail, and the skin surrounding the anus and the gluteal region. The efferent vessels end into the medial iliac LNs or lumbar trunks (Saar & Getty, 1982; Sugimura et al., 1958). ∼ Sacral LNs: found in relation to the middle sacral vessels. These nodes are inconstantly found in cats. When they are present, they drain lymph from the rectum, pelvic wall, tail, and hindlimbs. The efferent vessels drain into the medial and internal iliac LNs (Bezuidenhout, 2013; Saar & Getty, 1982; Sugimura et al., 1958; Tompkins, 1993).

• Inguinofemoral lymph center (lymphocentrum inguinofemorale)

Formerly known as superficial inguinal lymph center, its designation was changed by the NAV because some structures are not exactly located in the inguinal area (NAV, 2012). This lymph center includes the superficial inguinal, subiliac, and caudal epigastric lymph nodes (NAV, 2012). ∼ Superficial inguinal LNs: these nodes can be called mammary LNs in females and scrotal LNs in males. Their location is at the junction of the caudal superficial epigastric vessels with the external pudendal vessels in the inguinal region. One or two nodes can be found. The afferent vessels arise from the caudal abdominal wall, caudal epigastric LNs, skin from the gluteus and anal regions, and also the medial aspect of the hindlimb. In females, these LNs receive lymph from the caudal mammary glands and, in males, from the penis, prepuce, scrotum, and testis. The efferent vessels drain into the

LITERATURE REVIEW 43

medial iliac LNs (Bezuidenhout, 2013; NAV, 2012; Saar & Getty, 1982). ∼ Subiliac LNs: according to Sugimura et al. (1956), they are rarely found. When present, they are located at the mid-cranial border of the sartorius muscle, along to one of the caudal branches of the deep circumflex iliac vessels. Their afferent vessels arise from the gluteal and lumbar regions and the skin of the hindlimb. Efferent vessels drain into the medial iliac LNs (Saar & Getty, 1982; Sugimura et al., 1956). ∼ Caudal epigastric LNs: only present in the cat and the rabbit, these LNs are located along the caudal superficial epigastric vein more cranially than the superficial inguinal LNs. They are not always present. The afferent vessels arise from the abdominal wall and in females from the mammary glands of the middle abdomen. The lymph drain into the inguinal superficial LNs (Saar & Getty, 1982).

• Ischiatic lymph center (lymphcentrum ischiadicum)

It is only represented by the ischiatic LN. It is located along the caudal gluteal vessels, medially to the gluteofemoralis muscle in the ischiatic region. Afferent vessels arise from the tail, anal region, hindlimb, and popliteal LNs. Efferent vessels drain into the medial iliac LNs (Saar & Getty, 1982; Tompkins, 1993).

e. Hindlimb lymph centers

The lymph centers in this region drain lymph from the hindlimb as well as from the ventro-caudal portion of the abdominal wall. Getty et al. (1982) reported two lymph centers for the pelvic limb, the iliofemoral and the popliteal lymph centers.

44 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

• Iliofemoral lymph center (lymphocentrum iliofemorale)

It consists of the iliofemoral LN, but it is not always present. This node is located at the junction of the external iliac and deep femoral veins just dorsal to the inguinal canal. It drains lymph from the distal lateral surface of the hindlimb, the ventro-caudal portion of the abdominal wall and also from the popliteal and superficial inguinal LNs. The efferent vessels end into the medial iliac LNs (NAV, 2012; Saar & Getty, 1982).

• Popliteal lymph center (lymphocentrum popliteum)

In cats, as in dogs, this center is represented by the superficial popliteal LN. It is located subcutaneously in the popliteal fossa between the semitendinosus and biceps femoris muscles, at the caudo-proximal aspect of the gastrocnemius muscle. Normally there is only one node in each hindlimb. The afferent vessels arise from the distal part of the hindlimb. The efferent vessels drain into the iliofemoral and ischiatic LNs when they are present and/or the medial iliac LNs (NAV, 2012; Saar & Getty, 1982).

2.2. Ultrasonography of the lymphatic system

Ultrasound (US) is the term that describes sound waves of frequencies exceeding the range of human hearing and their propagation in a medium (Bushberg, Seibert, Leidholdt, & Boone, 2002b). In diagnostic US, a pulse of ultrasound waves is directed into the body (Kealy, McAllister, & Graham, 2011b). Interaction of the waves with the tissue’s acoustic impedance produces an interpretable gray scale image of the organs.

LITERATURE REVIEW 45

2.2.1. Physical principles

Sound travels in waves that are able to carry information from one location to another (Drost, 2013). The way the waves interact with the tissues and transmit the information to the computer depends on the physic characteristics detailed below.

a. Propagation of sound

It refers to the way that the physical properties of a sound wave travel in the tissues. The sound wave is mechanical energy traveling through a continuous, elastic medium which causes the compression and rarefaction of the “particles” that compose it (Bushberg et al, 2002b). Compression is an increase of pressure that causes deformation in the tissue (Bushberg et al., 2002b; Mattoon & Nyland, 2015). Rarefaction follows compression, when the pressure is reduced the compressed particles transfer their energy to the adjacent one, this reduces the local pressure amplitude (Bushberg et al., 2002b). The velocity of propagation of the sound is affected by the tissue properties, mainly the tissue’s resistance to compression, that includes the density and the elasticity (Mattoon & Nyland, 2015) and that increases in stiff tissues and decreases in high density tissues. Table 1 shows a list of the velocity of sound in different tissues according to their density.

Table 1. Velocity of sound and density from different materials and tissues MATERIAL Density (kg/m2) Velocity (m/s) Air 1.2 331 Lung 300 600 Fat 924 1450 Water 1000 1480 Soft tissue 1050 1540 Kidney 1041 1561 Blood 1058 1570 Liver 1061 1549 Muscle 1068 1585 Skull bone 1912 4080 PZT1 7500 4000 1PZT: lead zirconate titanate (Modified from Bushberg et al., 2002)

46 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

b. Frequency

Defined as the number of times a wave oscillates through a cycle per second and it is expressed in hertz (Hz). It is inversely related to wavelength; the higher the frequency, the shorter is the wavelength and vice versa. This relation is very important for the image resolution. A short wavelength (high frequency) produces a better resolution. The formula:

Velocity (m/sec) = frequency (cycles/sec) × wavelength (m)

shows the relationship between frequency, velocity, and wavelength (Bushberg et al., 2002b; Drost, 2013; Widmer, Mattoon, & Nyland, 2015).

c. Wavelength

The distance between compressions or rarefactions, or the distance traveled during one cycle. It is expressed in millimeters or micrometers (Bushberg et al., 2002b; Mattoon & Nyland, 2015).

d. Reflection

It is when a sound wave interacts with the tissue and is sent back toward the transducer and is the basis of ultrasound image. The reflection is affected by the angle in which the wave hits the tissue. A 90-degree angle is wanted for a better image quality (Drost, 2013; Mattoon & Nyland, 2015).

e. Refraction

It refers to the change in the direction of the US energy transmitted at the tissues boundary if the beam angle is not perpendicular to it, this might produce a change in the speed of the US wave; however, the frequency does not change. When a curved structure is being scanned, the combination of refraction and reflection contributes to the formation of an artifact called edge

LITERATURE REVIEW 47

shadowing laterally and distally to the structure (Bushberg et al., 2002b; Mattoon & Nyland, 2015).

f. Acoustic impedance

It is the product of the physical density of a tissue and the sound speed within the tissue. The amplitude of a returning echo is proportional to the difference in acoustic impedance between two tissues as the sound passes through their interface. Acoustic impedance can be calculated using the following formula:

Acoustic Impedance = Velocity × Tissue density

Sound velocity in soft tissues is considered constant (1540m/s), therefore the acoustic impedance basically depends on the tissue density (Drost, 2013; Mattoon & Nyland, 2015).

2.2.2. Modes of image display

a. B-mode (Brightness mode, B-scan or gray scale)

It is the mode most frequently used for regular abdominal US as well as for the examination of other soft tissue regions. The image is made by a collection of dots that are presented in a black background (Drost, 2013). The returning echoes are digitalized and converted into various intensities of brightness in two dimensions on a gray-scale format and are displayed on a monitor (Kealy et al., 2011b). The brightness of the dot is proportional to the amplitude of the returned echo and the depth of the structure that returned the echo determines the position of the dots relative to the position of the transducer (Drost, 2013; Mattoon & Nyland, 2015).

48 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

2.2.3. Instruments

The machines used in diagnostic US have a pulser that applies timed high- voltage pulses to the piezoelectric crystals within the transducer to generate sound waves. The crystals are made of lead-zirconate-titanate (PZT), a ceramic material that produces a mechanic vibration when an electric pulse is applied (Bushberg et al., 2002b). The transducer or probe is the transmitter and receiver of sound waves to and from the body, respectively, and is key to the image formation. The electric pulse is transformed in a sound wave that is reflected by the tissues and sent back to the transducer that transforms it in an electric pulse. Images are formed on the computer screen. The US machines also have controls that allow the operator to change or adjust contrast, frequency, and focal points, among others (Drost, 2013; Kealy et al., 2011b; Mattoon & Nyland, 2015).

2.2.4. Ultrasonography of the normal lymph nodes

The appearance and ultrasonographic features of LNs have been studied in both human and veterinary medicine (Nyman & O’Brien, 2007; Wunderbaldinger, 2006). The evaluation of LNs size, shape, margins, echogenicity, echopattern (echotexture), acoustic transmission, vascular flow (presence and distribution), and measurement of vascular flow indices have been reported as important features in the determination of normality or abnormality of LNs in veterinary literature (Llabres-Diaz, 2004; Nyman, Kristensen, Skovgaard, & McEvoy, 2005; Nyman & O’Brien, 2007). In human medicine, LNs are described as low echogenic oval and round structures with a clear visible hilum containing the central vessels (Wunderbaldinger, 2006).

The LNs size can be assessed by a short-to-long-axis (S/L) ratio that helps to guide the diagnosis of lymphadenomegaly. A ratio ≤0.5 has been reported to be normal in medial iliac LNs in dogs (D’Anjou, 2008; Llabrés-Díaz, 2004).

LITERATURE REVIEW 49

Descriptions about the normal shape of LNs includes fusiform to oval (D’Anjou, 2008), slender (Nyman & O’Brien, 2007), round or oval to more elongate and fusiform (Mattoon, Berry, & Nyland, 2015). Normal margins have been described as smooth (D’Anjou, 2008), sharp (Nyman & O’Brien, 2007), or variable contour (Kealy, McAllister, & Graham, 2011a).

The normal echogenicity of the LNs have been reported to be nearly anechoic (Mattoon et al., 2015), hypoechoic (Kealy et al., 2011a), slightly hypoechoic, relatively isoechoic or even echogenic to adjacent fatty tissues (D’Anjou, 2008; Mattoon et al., 2015; Nyman & O’Brien, 2007). The normal echopattern has been described as homogeneous (Mattoon et al., 2015) or mildly heterogeneous (Nemanic & Nelson, 2012).

Superficial LNs are better assessed using a high frequency probe (7.5 – 13MHz or higher), considering that these structures are small in size and very often are loose in the surrounding fat. There are few reports about the technique to perform an ultrasonographic evaluation of the superficial LNs in the cat. Literature giving description of the features for superficial LNs is only available for the ultrasonographic evaluation of the retropharyngeal (Nemanic & Nelson, 2012) and popliteal LNs (Lee et al., 2012).

Abdominal LNs are not typically easy to assess. An important factor that should be taken into account is the body condition; as the body condition increases, the difficulty in identifying a normal LN also increases. The attenuation of the ultrasound beam caused by the fatty tissue that surrounds the LN could be an explanation for this (Mattoon et al., 2015).

Some other factors have to be considered when the thoracic and the abdominal LNs need to be assessed. The presence of air in the lungs and in the intestinal loops makes difficult the visualization of LNs due to the acoustic shadowing that it produces (Mattoon et al., 2015; Nyman & O’Brien, 2007).

50 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

An evaluation of the whole lymphatic system is never performed in a single patient because, in a clinical situation, a regional evaluation is usually required.

2.2.5. Ultrasonographic features of abnormal lymph nodes

Commonly, a malignant LN is rounded in shape with hypoechoic parenchyma (Nyman & O’Brien, 2007). However, if hemorrhage, coagulative necrosis or mineralization is present, the appearance of the LN is more heterogeneous (D’Anjou, 2008). Ultrasonographically, heterogeneity of LNs has been described as a significant finding associated with malignancy in dogs. However in cats, a statistical association of abdominal LNs heterogeneity with malignancy was not found (Kinns & Mai, 2007).

Lymphadenopathy is associated with reactive, inflammatory, and neoplastic processes. Ultrasonographic features between normal, reactive, and malignant (metastatic) lymphadenopathy overlap (Nyman & O’Brien, 2007). Massively enlarged LNs showing heterogenic areas may be related with benign or malignant infiltration and can be associated with a pyogranulomatous disease or neoplasia (D’Anjou, 2008).

In humans, contrast enhance ultrasound (CEUS) is preferred because allows an accurate noninvasive lymph node assessment and cancer staging (Wunderbaldinger, 2006). In veterinary medicine, descriptions of the applications and uses of this modality are limited. So far, the study of CEUS in dogs have lead to reports that advice its use for detection and characterization of focal lesions in the liver (Nakamura et al., 2010) and spleen (Rossi, Leone, Vignoli, Laddaga, & Terragni, 2008). A study has shown that contrast harmonic ultrasound accurately depicted angioarchitechture within lymphomatous nodes in dogs; this additional information can be used in determining the presence of malignant vascular

LITERATURE REVIEW 51

characteristics of LNs (Salwei, O’Brien, & Matheson, 2005). The use and applications of CEUS in cats are also limited. A recent study about normal vascular pattern and appearance of normal abdominal organ and mesenteric LNs concluded that CEUS can be used to estimate organ perfusion like in other species (Leinonen et al., 2010). Studies showing the utility of CEUS in differentiating vascular patterns in lymphadenopathies are lacking.

2.3. Computed tomography of the lymphatic system

Computed tomography (CT) was developed in the 70’s as a computer system that allowed the acquisition and reconstruction of images from a specific region of the patient avoiding the superimposition among organs. CT technology was invented by Godfrey Hounsfield and Allan Cormack based on the mathematic principles of Radon that are also used in conventional radiography (Bushberg, Seibert, Leidholdt, & Boone, 2002a; D’Anjou, 2013). The first CT scans had long acquisition times, but with the advance of technology a full body CT scan can be nowadays done in seconds. In veterinary medicine, the first CT scans were performed in the late 80’s and patients had to be taken to a human hospital to be scanned. The first CT scan for veterinary use was installed at the École Nationale Vétérinaire d’Alfort in Paris in 1989, allowing the beginning of CT diagnostic imaging in veterinary medicine (Schwarz & Saunders, 2011).

2.3.1. CT principles

Radon’s mathematic principles proved that an image of an unknown object could be reproduced if an infinite number of projections through the object had been acquired. This means that the CT image is a picture of a slab (or slice) of the patient’s anatomy (Bushberg et al., 2002a).

52 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

2.3.2. CT scans components

Basically all CT machines are conformed by the same components, and their designs vary depending on the manufacturer.

a. Gantry

Is a doughnut-shape ring that contains the rotating X-ray tube, the detectors array, the collimator and the hardware necessary to generate radiation and to detect the attenuated X-rays that had passed through the patient and that later allow the reconstruction of the images. The gantry has a hole in the center that allows the patient table to slide in and out according to the scanning protocols (Bushberg et al., 2002a; Saunders & Ohlerth, 2011). The gantry is one of the components that has had more modifications along the time helping in reducing the acquisition times, artifacts, and image quality (Bushberg et al., 2002a).

b. X-ray tube

A vacuum cage enclosing an X-ray tube, designed to resist the demanding protocols of a CT scan. One of the major differences with a conventional radiography tube is the fan beam geometry that it uses, producing a X-ray beam in a fan shape (Bushberg et al., 2002a; Saunders & Ohlerth, 2011).

c. Collimator

It is located between the X-ray source and the patient (pre-patient collimators) and between the patient and the detectors (post-patient collimators). The collimators are useful to ensure good image quality and to reduce unnecessary radiation doses to the patient. Another important characteristic is that the adjusted opening allows the selection of a slice width and the size and position of the focal spot (Saunders & Ohlerth, 2011).

LITERATURE REVIEW 53

d. Detector system

Another important component that is located in the gantry in an opposite position to the X-ray tube. In the 3rd generation scanners, a detector array rotates with the X-ray tube inside the gantry. It is long enough to receive the radiation from the fan beam, including the entire width of the patient; this conformation is widely used in most scanners. In the 4th generation scanners a detector array is disposed around the gantry ring. Nowadays, the detectors are made from ceramic solid-state materials because they have a better X- ray absorption efficiency. The detector has two main functions: 1) the reception of the incident X-ray photons, which have passed through the patient and 2) the transformation of the photon into a electrical sign, that is then amplified and converted from analog to a digital form (Bushberg et al., 2002a; Saunders & Ohlerth, 2011).

e. Patient table

It is made of carbonate fiber that will not produce artifacts in the image when scanned. Usually it is fixed in the ground but the height can be adjusted. The table slides in and out of the gantry opening, allowing the irradiation of a specific region of the patient that is the objective of the examination (Bushberg et al., 2002a).

f. Work station

It is normally located in a separate room covered with lead shields that provide protection against the radiation to the operators. The main computer to program protocols, to reconstruct, and to display the images is also located in this room (Saunders & Ohlerth, 2011).

54 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

2.3.3. Image acquisition principles

They involve all the steps that have to be taken into account when scanning. Body part selection is the first part of the process because it will determine the algorithms in which the image will be reconstructed. Normally CT scans use protocols that include the tube rotation time, mA, mAs and kV for different regions. An explanation of other concepts that need to be considered is given below.

a. Scan field of view (SFOV)

The x-ray fan beam circles the gantry while emitting radiation. The image is reconstructed from the data of the area in the gantry where there is a complete overlap of the x-rays fan beam projections. The region of the body that is to be scanned should not be outside the SFOV (Bushberg et al., 2002a; Schwarz & O’Brien, 2011).

b. Display field of view (DFOV)

Area of the SFOV from which the image is reconstructed (Bushberg et al., 2002a; Schwarz & O’Brien, 2011).

c. Slice thickness

Refers to the tissue thickness that will be used to attenuate the x-ray beam and from which the slice is reconstructed. The collimator and detectors array play an important role in the determination of slice thickness. The different tissues contained in a slice influence the quality of the image. An increase in the slice thickness can produce a better visualization of some structures, but can also increase the possibility of acquiring a blurry image (Schwarz & O’Brien, 2011).

LITERATURE REVIEW 55

d. Slice interval

It refers to the interval in which the images are acquired in sequential mode (axial); it can be programmed to be the same distance of the slice thickness to ensure continuous image acquisition. In helical mode, data are obtained continuously and this interval is not used (Schwarz & O’Brien, 2011).

e. Helical pitch

It is the expression of the relationship between the table increment during one full gantry rotation and the slice thickness. It is directly proportional to image blur, meaning that a high CT pitched scan produce a highly blurry image. A pitch of zero will result in a complete ring of data (sequential scan). A pitch of 1 results in a stretching of the helix by one degree that after one rotation the table will have moved by one slice or collimator width. The image reconstruction will be done with mathematical interpolation of the data contained in other parts of the helix. A very high pitch makes the mathematical interpolation (guesses) incorrect which produces a blurry image (Schwarz & O’Brien, 2011).

2.3.4. Image reconstruction

Determined by the different type of tissue that is contained in a slice thickness. The slice is divided in a matrix that is conformed by pixels showing the average attenuation of each voxel. In veterinary, a matrix of 512x512 is commonly used.

a. Voxel (Volume element)

Volume area obtained multiplying the slice thickness by the pixel area (Bushberg et al., 2002a).

56 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

b. Pixel (Picture element)

It is a 2D square that shows a shade from a gray scale (Hounsfield gray scale) that is obtained by averaging the attenuation values of all the tissues that are contained in the voxel. The pixel is the face of the voxel (Bushberg et al., 2002a).

c. Hounsfield Units (HU)

Developed by Geoffrey Hounsfield, these units correspond to the numeric values that give to each pixel a shade from a gray scale. The scale has “0” as a central value given to the attenuation of the water and is expressed as a specific tone of gray. Structures with more attenuating characteristics are represented with brighter gray colors until white is reached. Structures with less attenuating characteristics are represented in darker gray colors until black is reached. The range of the HU is -1000 to +3095 producing a total of 4096 shades of gray. However, the human eye cannot differentiate more than 90 shades of gray. To solve this problem, the image gray scale must be adjusted using the window width and level (Bushberg et al., 2002a; D’Anjou, 2013; Saunders & Ohlerth, 2011). Table 2 shows the reported HU for different organic tissues.

d. Window width

Range of HU numbers (minimum to maximum) displayed in the image. This determines image contrast.

e. Window level

It is the central HU number of the chosen window width. This determines image brightness. The window level should match the density level of the organ of interest.

LITERATURE REVIEW 57

The combination of the window width and level is responsible for the image display for the different windows: bone, soft tissue, lung, brain (D’Anjou, 2013; Saunders & Ohlerth, 2011).

Table 2. Hounsfield Units (HU) values for organic tissues and fluids Tissue HU standard value Compact bone >250 Trabecular bone 50 – 300 Coagulated blood 70 – 90 Thyroid gland 60 – 80 Liver 50 – 70 Whole blood 50 – 60 Brain gray matter 37 – 41 Muscle 35 – 50 Pancreas 30 – 50 Kidney 20 – 40 Brain white matter 20 – 34 Plasma 27 ± 2 Exudates (>30 g protein/L) >18 ± 2 Transudates (<30 g protein/L) <18 ± 2 Cerebrospinal fluid 5 – 10 Fat -80 to -100 Lung -950 to -550 Modified from Ohlerth & Scharf (2007)

2.3.5. Multiplanar reconstruction of images

CT scans normally reproduce images in an anatomical transverse plane. However, any other anatomic plane can be reconstructed with the data recovered by the scanner. Another type of reconstruction is the volume rendering that reproduce a 3D image of the patient (Saunders & Ohlerth, 2011).

2.3.6. Contrast studies

Iodinated contrast media are commonly used and provide information about blood flow, perfusion of tissues, and integrity of natural barriers. These media

58 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

are administrated with a catheter into the blood stream, usually a peripheral vein; however, intra arterial, intrathecal, and intra articular administration can also be performed depending on the procedure. The contrast agent is used with the purpose of intensifying the visualization of structures (mainly soft tissues) and its vascular permeability (D’Anjou, 2013).

2.3.7. Computed tomography of normal lymph nodes

In human medicine, CT is widely available and the systematic visualization of all LNs location, and combined evaluation of other lesions and surrounding structures is possible and has been done for the detection and follow up of diseases (Wunderbaldinger, 2006).

In veterinary medicine, the assessment of LNs also has an important role in grading or establishing the severity of the pathologies that can affect the patient. Only a few studies regarding the normal appearance of the LNs in the cat are available in the literature. Some studies have assessed the retropharyngeal (Nemanic & Nelson, 2012), the cranial mediastinal, and the tracheobronchial LNs (Henninger, 2003) in order to describe their appearance. A mildly heterogeneous, 20.7mm length X 13.1mm rostral height X 4.7mm rostral width of a medial retropharyngeal LN was reported as normal (Nemanic & Nelson, 2012). In the study for the cranial mediastinal and the tracheobronchial LNs, these were not identified and the authors report that the reason was they were not enlarged (Henninger, 2003).

In dogs, abdominal LNs have been described as commonly elongated in shape, homogeneously attenuating structures before (37HU, range: 20 – 52) and following (109HU, range: 36 – 223) the intravenous administration of contrast medium. Some LNs were slightly irregular or relatively more hyperattenuating in the periphery than centrally before and after contrast administration (Beukers, Vilaplana Grosso, & Voorhout, 2013).

LITERATURE REVIEW 59

There are no studies about the normal appearance of cervical and abdominal LNs in the cat.

2.3.8. Computed tomography of abnormal lymph nodes in the cat

In human medicine, the LNs are considered malignant in CT images if they present a short diameter of more than 10mm, a S/L ratio >0.5 or a L/S (long- to-short-axis) ratio <2, detectable extra nodal tumor spread and/or necrotic or cystic changes and/or heterogeneous contrast enhancement, or are clearly increased in number (Mohseni et al., 2014; Steinkamp et al., 1995; Steinkamp, Hosten, Richter, Schedel, & Felix, 1994; Tohnosu, Onoda, & Isono, 1989; Vassallo, Wernecke, Roos, & Peters, 1992). However, the low sensitivity in detecting LNs with an histologic diagnosis compatible with metastatic infiltration has lead to the combination of CT with other techniques like PET (positron emission tomography), improving the accuracy in the assessment of LNs and cancer staging (Wunderbaldinger, 2006).

In cats, there are studies about the characteristics of different diseases and some of them include a brief description of one or more LNs that are involved. In a study about nasal polyps in cats, mandibular and medial retropharyngeal LNs presenting ellipsoidal shape, heterogeneous contrast enhancing, hyperattenuating foci centrally and peripherally, and a length of 29.4mm were considered suggestive of lymphadenopathy (Oliveira, O’Brien, Matheson, & Carrera, 2012). In a study about the CT findings in fungal rhinitis and sinusitis, 5/9 cats had images compatible with lymphadenopathy. In 1/5 cats, enlarged right mandibular and retropharyngeal LNs with heterogeneously contrast enhancing with central hypoattenuation were seen. A homogenously enhancing of both LNs was present in 1/5 cats. In 3/5 cats only retropharyngeal lymphadenopathy was identified. Heterogeneous contrast enhancement with central hypoattenuation was present in 2/3 cats. A homogeneous enhancement of the retropharyngeal LNs was seen in 1/3

60 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

cats. No indication of size and HU were reported in this study (Karnik, Reichle, Fischetti, & Goggin, 2009). In a study regarding the CT features of oral squamous cell carcinoma, the maximum width of the mandibular (4.1 +/- 1.9mm) and the medial retropharyngeal (5.3 +/- 1.5mm) LNs was measured on postcontrast images (Gendler, Lewis, Reetz, & Schwarz, 2008). Other studies about thoracic neoplastic disease (Henninger, 2003) did not describe CT features of the identified LNs (tracheobronchial).

There is no literature about the assessment of pathologic abdominal LNs in the cat.

3. OBJECTIVES

OBJECTIVES 63

3. Objectives

§ To assess the CT and US ability to identify the lymph nodes of healthy cats comparing the imaging features with measurements from normal anatomic values.

§ To characterize the lymph nodes in diseased cats using CT and US.

§ To assess the ability of each imaging technique (CT and US) to discriminate between neoplastic and inflammatory changes in the lymph nodes.

4. STUDIES

4.1. Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb.

STUDIES 69

Abstract

The assessment of the lymph nodes (LNs) is key in staging cancer patient. Descriptions about the normal computed tomography (CT) and ultrasound (US) features of feline LNs are limited in the veterinary literature. The purposes of this study were (i) to compare the size of the head, neck, forelimb, and thorax LNs obtained with US and CT in a group of healthy adult cats with measurements obtained from an anatomic study and (ii) to describe the US and CT features of these LNs in healthy adult cats. An anatomic study in 6 cadavers and an imaging study (CT and US) in 30 healthy cats were performed. The frequency of identification of the lymph centers varied among techniques and also, individually. The mandibular LNs (MnLNs) were the only ones identified in 100% of the cats both in the anatomic and the imaging studies. The medial retropharyngeal LNs (MRLNs) were identified in 100% of the cats using CT and US. The deep cervical LNs were not visualized in the cadavers. The cranial mediastinal and tracheobronchial LNs were not visualized using US. Length, width, and height were measured and compared among techniques. The length of the LNs on CT was higher when compared with US and anatomic lengths. The highest differences were in the MRLNs; the length in CT was 6mm larger than the length in US and anatomy. Ultrasonographically, LNs were relatively wider than on CT and anatomy. The highest differences were presented for the MRLNs and superficial cervical LNs (SCLNs). The height was the most variable measurement among techniques showing the highest statistically significant differences among them. The LNs identified with CT were most frequently isoattenuating or slightly hypoattenuating to the surrounding musculature, with homogeneous contrast enhancement. In US, most LNs were isoechoic or hypoechoic to surrounding fat tissue. The majority of LNs were fusiform or rounded. Nevertheless, the MRLNs, the dorsal SCLNs, and the accessory axillary LNs (AAxLNs) had miscellaneous shape. Some conditions that improved the visualization of the LNs were identified; the fat tissue provided a good contrast to differentiate the LNs from the other soft tissues. In thin cats, the CT contrast studies helped in the differentiation of the LNs. The measurements and features reported are proposed as reference values.

70 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

Introduction

Lymph nodes (LNs) characterization is important in the diagnosis and prognosis of neoplastic and infectious diseases (Nemanic, Hollars, Nelson, & Bobe, 2015; Nyman, Kristensen, Skovgaard, & McEvoy, 2005; Nyman & O’Brien, 2007). There is limited information regarding the imaging features of the lymph centers in the cat. Normal values for size, appearance, and ability to depict many of them using computed tomography (CT) and ultrasonography (US) are needed. The lymph centers (LCs) for the head, neck, thorax, and forelimb are reported in the Nomina Anatomica Veterinaria (NAV). The parotid, mandibular, and retropharyngeal LCs are described in the head; the superficial and deep cervical LCs in the neck; and the axillary LCs in the forelimb. In the thorax, four LCs exist, divided into parietal and visceral LCs. The parietal LCs include those located on the inner side of the thoracic wall: the dorsal and ventral thoracic LCs. The visceral LCs include those located within the pleural and mediastinal spaces: the mediastinal and bronchial LCs (Bezuidenhout, 2013; NAV, 2012; Tompkins, 1993). The normal length of the LNs in the cat has been reported in the anatomy literature mainly based on the studies made in Japan in the 50’s (Saar & Getty, 1982; Sugimura, Kudo, & Takahata, 1955; Sugimura, Kudo, & Takahata, 1959). However, the normal width and height were not reported in those anatomic references. Description of normal US and CT features of the lymph nodes in the head, neck, forelimb, and thorax of the cat is lacking in the current literature. The aims of this study were: (i) to compare the size of the head, neck, forelimb, and thorax lymph nodes obtained with US and CT in a group of healthy adult cats with measurements obtained from an anatomic study; and (ii) to describe the US and CT features of these lymph nodes in healthy adult cats.

STUDIES 71

Materials and methods

The ethical committee of the Universitat Autònoma de Barcelona approved this study; reference number CEAAH 2255 of September 2013. The owner consent for all the patients and cadavers included in the study was also obtained.

Anatomical study

Animals

Feline cadavers referred for necropsy to the pathology department of the Universitat Autònoma de Barcelona, within 24 hours of dead, were prospectively included from January 2013 to June 2015. Using a 24 scalpel blade, the skin was removed making an incision from the mandibular symphysis to the xiphoid process of the sternum. Then, the incision continued laterally and dorsally in the direction of the armpit and caudal to the scapula. A second incision was made around the elbow joint and in the medial aspect of the brachium to connect it with the first incision. The skin was then separated from the platysma and removed. Lymph nodes were searched at their anatomical location, mainly following the veins. The axillary lymph center was searched following the axillary and lateral thoracic vessels. Then, the costochondral joints were carefully cut in order to expose the thoracic cavity. The diaphragm was separated from the thoracic wall and blunt dissection was used to identify the lymph nodes inside the thorax. Postmortem coloring procedures for the lymphatic system were not performed because the cats were evaluated after several hours of dead. The length, height, and width of each LN were measured using a manual dial caliper (Vernier 0 – 150mm/ 0.02 high precision). The length was defined as the largest dimension in the rostro/craniocaudal plane. Height was measured at the thickest point in the dorsoventral plane. Width was measured at the thickest point in the mediolateral plane. The number of lymph nodes per lymph center, as well as the anatomical landmarks, shape, and size was recorded.

72 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

Imaging study

Animals

Healthy cats older than 1 year of age were recruited at the Fundació Hospital Clinic Veterinari of the Universitat Autònoma de Barcelona (FHCV-UAB) from staff, students, and hospital clients. The animals were considered healthy based on physical exam, biochemical profile [calcium, glucose, potassium, total proteins, alanine-amino-transferase (ALT), gamma-glutamyl-transferase (GGT), cholesterol, urea, creatinine] and complete blood count. A SNAP® test to rule out the presence of feline immunodeficiency virus (FIV) antibodies and feline leukemia virus (FeLV) antigens, and a PCR test to rule out the presence of Bartonella sp were also performed.

Computed Tomography

The animals were sedated with an intramuscular administration of midazolam1 (0.2mg/kg), butorphanol2 (0.4mg/kg), and ketamine3 (5mg/kg). Anesthesia 4 was induced with Isoflurane 5% dosage 100% O2 at 4L/min and maintained

with Isoflurane 1.5 – 2% in 100% of O2 at 2L/min. The patients were positioned on the CT table in dorsal recumbency with the forelimbs and hindlimbs outstretched at the sides. A whole body scan was performed. Acquisitions were done in soft tissue algorithm, before and after the intravenous administration of 600mg/kg of Iopromide5 (300mgI/ml) or Iopamidol6 (300mgI/ml) in the cephalic vein. Scans were performed in a 16 slices helical CT-scanner7 with a slice thickness of 0.625mm, interval thickness of 0.625mm, collimation pitch of 1.25mm, 120kV, 50 - 90mA, and a matrix of 512 x 512.

1 Midazolam 15mg/3ml, Normon, Spain 2 Torbugesic 10 mg/ml, Zoetis, Alcobendas (Madrid), Spain 3 Imalgene 100 mg/ml, Merial, Barcelona, Spain 4 Isoflurane, Abbott Laboratories, Berkshire, UK 5 Ultravist® 300mg/ml, Bayer pharma AG, Berlin, Germany. 6 Scanlux® 300mg/ml, Sanochemia pharmazeutika, Neufled/Leitha, Austria. 7 General Electric® Brivo CT 385.

STUDIES 73

Image analysis: For each lymph node identified, CT characteristics and measurements were performed following previously described methods for the measurement and comparison of CT images (Nemanic & Nelson, 2012). All data were recorded for further analysis using an image archiving and communication system software.8 Measurements were performed following previous descriptions for dogs (Beukers, Vilaplana Grosso, & Voorhout, 2013; Nemanic & Nelson, 2012). The length was determined using two previously reported methods; (1) Calculated CT LN length: multiplying the slice thickness by the number of transverse images that contained the lymph node; and (2) Measured CT LN length: in a multiplanar reconstruction (MPR) to generate a sagittal image of the LN at its maximal dimension from rostral/cranial to caudal; an electronic caliper was placed from the rostral/cranial to the caudal border to measure the length of the lymph nodes. Height and width were measured in transverse images at the rostral/cranial, middle, and caudal aspects of each lymph node. Height was defined as the distance from the ventral to the dorsal border in each position and width was defined as the distance from the medial to the lateral border in each position. The highest values were used for the statistical analysis. The short / long (S/L) axis ratio was calculated using the higher value of height divided by the value of length obtained in the multiplanar reconstruction and this was calculated for each LN. The shape of the lymph nodes was classified as rounded, elongated, or miscellaneous as previously reported by Beukers et al. (2013) and Nyman, Kristensen, Skovgaard, & McEvoy (2005). A lymph node was defined as rounded when the short axis/long axis ratio was >0.5. A short axis/long axis ratio ≤0.5 was used to classify a lymph node as elongated. Lymph nodes with a multilobular structure that did not fit the ratio were classified as miscellaneous. The attenuation (Hounsfield units) was determined by placing a circular/oval region of interest (ROI) of 2-4mm2 over the same rostral/cranial, middle and caudal transverse slice where width and height measurements were performed. In small LN, ROIs were made as large as possible inside the lymph node margins. Attenuation measurements were performed before and following the administration of contrast medium. Mean

8 Centricity PACS-IW, GE healthcare.

74 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

values for attenuation pre- and postcontrast were calculated using the three previously obtained measures. As in previously reported studies, lymph nodes attenuation was compared with the surrounding muscles and was classified as isoattenuating (same attenuation), slightly hypoattenuating (minimal homogeneous decrease in attenuation), hypoattenuating (marked homogeneous decrease in attenuation), hyperattenuating (homogeneous increase in attenuation) and heterogeneous (single or multiple areas of different attenuation within the LN). Following the administration of contrast medium the attenuation was classified as homogeneous, mildly heterogeneous (small, multiple areas of different contrast enhancement), heterogeneous (large, multiple areas of different contrast enhancement), and peripheral enhancement (contrast enhancement in a ring-like distribution with a hypoattenuating center). These features were used to establish CT reference values for feline lymph nodes.

Ultrasonography

Immediately after the CT scan and with maintained anesthesia, an ultrasound scan was performed to each animal. The hair of the ventral aspect of the neck was clipped from the intermandibular region to the mid-cervical region and extended laterally to include the parotid region (at the level of the mandibular angle). The axillary region extending ventrally to the sternum and the cranial aspect of the shoulders extending dorsally following the scapula at both sides were also clipped. The animals were positioned in dorsal recumbency with the neck extended and the forelimbs slightly extended to the sides. Right and left lateral recumbency was also used when the dorsal cervical LNs were assessed. Examinations were performed using an Esaote Mylab70 Xvision® machine with a 4 – 13MHz frequency linear transducer. Technical settings were adjusted to improve and obtain the optimal images of the LN in all the animals. Acoustic coupling gel9 was generously applied to ensure an adequate skin-transducer contact. Sagittal and transverse images of each lymph node were recorded.

9 Transonic gel®, Telic, Barcelona, Spain

STUDIES 75

Image analysis: for the sagittal plane, the transducer was placed with the guide pointing rostral/cranial, parallel (or slightly oblique) to the spine, and an image including the largest measurement of the LN was recorded. Measurements in this image were performed using an electronic caliper from the rostral/cranial to the caudal border (long axis) and was defined as the length of the LN. A second measurement was performed in the same image perpendicularly to the first measure at the point of maximum thickness (ventral to dorsal) and was defined as the height (short axis) of the LN. With these two measurements, the ratio short axis/long axis was calculated. For the transverse plane, the transducer was rotated 90º with the guide towards the right side of the patient selecting an image that contained the largest portion of the lymph node. A measurement was performed from medial to lateral and was considered as the LN width. For each lymph node, echogenicity was recorded as hypoechoic, isoechoic, hyperechoic, or heterogeneous when compared to surrounding fat tissue. The presence of a hyperechoic central line that corresponds with the hilus was also recorded. The shape of each lymph node was evaluated following the same criteria as in CT. Margins were defined as smooth or irregular.

Statistics

Data were digitalized using Excel (2010)10. Statistical analyses were performed using the free available statistics software R (2015)11. The frequency of LNs identification, mean and SD of attenuation values pre and postcontrast administration, echogenicity, and the mean and SD of LNs measurements was calculated. Wilcoxon Signed Rank Test was used to compare the pair distribution between the calculated length and the length obtained with the multiplanar reconstruction (length MPR) of the LNs on CT images. After this, the length MPR was used in the pair comparison with the US. The rest of the LNs measurements (width and height) between TC and US were also compared with Wilcoxon Signed Rank Test. Mann-Whitney U test was used to compare the pair distribution of the LN measurements

10 Microsoft office Excel, 2010 11 R versión 3.2.3 (2015-12-10). Copyright © 2015, the R foundation for statistical computing.

76 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

(length MPR, width, and height) between TC and anatomy, and between US and anatomy. Each measurement was compared individually for each lymph center and not for the whole sample of identified LNs (No Post-Hoc corrections were used). A P value <0.05 was considered statistically significant.

Results

Animal description

Anatomic study: six feline cadavers were included. Causes of death were not related to neoplastic or inflammatory diseases according to the necropsy [heart failure (n=2), kidney failure (n=2), poisoning (n=1) and unknown (n=1)]. The average age was 6.8 years (range 1 – 16). Five cats were domestic shorthairs and one cat was a British longhair.

Imaging study: thirty cats were recruited. Age and weight averages were 3.7 years (range 1.5 – 17) and 4.4kg (SD 1.1), respectively. Twenty-nine cats were domestic shorthairs and 1 cat was a Persian. The group included 16.7% entire males (n=5), 20.0% neutered males (n=6), 30.0% entire females (n=9), and 33.3% neutered females (n=10). Biochemical determinations and complete blood count were within normal limits. All cats were negative for FIV/FeLV and Bartonella sp. tests. Table 1 shows the mean and SD for the length, width, and height, as well as the results of the statistical comparisons among techniques; table 2 shows the attenuation values (Hounsfield Units), and table 3 the ultrasonographic features.

Parotid lymph center (lymphocentrum parotideum)

In the anatomic study, the right and the left (PLNs) were identified in 3 cadavers with a total of 6 PLNs (3 right and 3 left). All were round or ovoid, located rostral to the parotid salivary gland, and in 1 cat both PLNs were covered by the salivary gland. The parotid salivary gland duct runs

STUDIES 77

ventrally to the LN and the temporal superficial vessels run dorsally. Fat tissue could be found around the LNs. On CT images, the right and left PLNs were identified in 8 cats, only the right PLN in 1 cat, and only the left PLN in 4 cats with a total of twenty-one LNs identified (9 right and 12 left) in 13/30 cats. The close contact with the parotid salivary gland made the differentiation between the LN and the glandular tissue challenging. However, the administration of contrast medium improved the visualization of these lymph nodes in some animals. When visible, they were ovoid in shape, homogeneously isoattenuating to the surrounding soft tissue in precontrast and showed homogeneous enhancement after contrast administration. On US images, only 1 parotid LN was identified in 1/30 cats. It appeared elongated, slightly hypoechoic compared to fat tissue, surrounded by a thin hyperechoic halo, and located rostral to the parotid salivary gland (Figure. 1).

Mandibular lymph center (lymphocentrum mandibulare)

Four LNs were identified in the 6 cadavers for this lymph center. They correspond to the right lateral (RL), right medial (RM), left lateral (LL) and left medial (LM) mandibular LNs (MnLNs). Each pair of LNs (medial and lateral) were located on each side of the mandibular angle; the linguofacial vein ran between the medial and lateral MnLNs, being the most important anatomic landmark. All the MnLNs showed an elongated shape. On the CT images, the four MnLNs were identified in the 30 cats with a total of 120 LNs. They were most frequently isoattenuating to the surrounding muscles, which made them challenging to identify in thin cats. After contrast administration, a homogeneous contrast enhancement was commonly observed. In the thin cats with isoattenuating LNs, the postcontrast images helped in the localization of these in the pre-contrast images. On the US images, the four MnLNs were also identified in the 30 cats with a total of 120 LNs. They appeared elongated in sagittal and ovoid in transverse planes. As for the parotid LNs, their appearance was most frequently slightly hypoechoic in comparison to the fat tissue, surrounded by a thin hyperechoic halo. A hyperechoic central line was present in a 7.7% of the MnLNs. When

78 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

Doppler was used, on transverse images, a blood vessel was easily identified between the MnLNs compatible with the linguofacial vein (Figure. 2).

Retropharyngeal lymph center (lymphocentrum retropharyngeum)

The right and the left medial retropharyngeal LNs (MRLNs) were identified in 3 cadavers, and only the left MRLN in 2 cats for a total of 8 MRLNs (3 right and 5 left) in 5 cats. The RMRLNs and LMRLNs were located medially to the respective mandibular salivary gland and the sternocephalicus muscle, ventrally to the caudal aspect of the tympanic bullae, longus colli muscle at the level of the first 2 cervical vertebrae, and laterally to each carotid sheath. These MRLNs were commonly oval in shape. The lateral retropharyngeal LNs were not found in the anatomic study. On the CT scans, both MRLNs were identified in the 30 cats for a total of 60 LNs. The MRLNs were frequently slightly hypoattenuating, followed by isoattenuating appearance and less frequently hypoattenuating. After contrast administration, the right and left MRLN frequently showed slightly heterogeneous contrast enhancement or homogeneous contrast enhancement, and less frequently heterogeneous contrast enhancement. Ultrasonographically, both MRLNs were also identified in the 30 cats for a total of 60 LNs. The RMRLNs were frequently hypoechoic, and less frequently isoechoic or heterogeneous to the surrounding tissues. Meanwhile, the LMRLN was seen mainly hypoechoic. A small proportion of isoechoic or heterogeneous LMRLN was also seen. Both MRLNs were commonly elongated (Figure. 3). In some of them a miscellaneous shape was observed. Also, in 13.3% of both right and left MRLN, a hyperechoic central line was identified. The lateral retropharyngeal LNs were not found in this study.

Superficial cervical lymph center (lymphocentrum cervicale superficiale)

In four cadavers, one dorsal and one ventral superficial cervical lymph nodes (DSCLNs & VSCLNs) were found on both sides. Additionally, only 1 left DSCLN was seen in 2 cadavers, for a total of 18 SCLNs in 6 cadavers. The

STUDIES 79

DSCLNs were located deep to the trapezium and omotransversarius muscles, related with the superficial cervical vessels and surrounded by fat tissue. The VSCLNs were located dorsally to the junction of the superficial cervical vein with the external jugular vein. The DSCLNs were commonly miscellaneous and the VSCLNs were frequently seen with an elongated shape. The DSCLNs and VSCLNs were commonly visualized on CT images. The most dorsally located DSCLN was found in all cats on the right side and in 29/30 cats on the left side. The most ventrally located DSCLN was found in 25/30 cats on the right side and in 24/30 cats on the left side. The VSCLNs were identified in 29/30 cats on the right side and in 25/30 cats on the left side. Both groups were more commonly observed with an isoattenuating or slightly hypoattenuating appearance. After contrast administration, they frequently showed a homogeneous contrast enhancement. On US examination, 2 DSCLNs and 1 VSCLNs were commonly identified on each side. The most dorsally located DSCLN was found in 9/30 (30%) cats on the right side and in 6/30 (20%) cats on the left side at the level of the cranial angle of the scapula. The most ventrally located DSCLN was found in 29/30 (96.7%) cats on the right side and in 25/30 (83.3%) cats on the left side at the level of the mid-cranial border of the scapula. The identification of all the SCLNs in both sides was not possible in all the cats. One dorsal and one ventral SCLNs on each side were found in 2 cats. The identification of one right and one left DSCLNs was possible in 18 cats. The shape of the most dorsally located DSCLNs was most frequently fusiform. The most ventrally located DSCLNs were miscellaneous (from ovoid to elongated) and were the biggest of the DSCLN. These LNs were slightly hypoechoic to the surrounding fat. The VSCLNs were seen in 2/30 (6.7%) cats on each side. They were elongated and showed iso to hypoechoic appearance with smooth margins. A hyperechoic central line was visible in 1.8% of the SCLNs (Figure. 4).

Deep cervical lymph center (lymphocentrum cervicale profundum)

The middle or caudal deep cervical LNs were not visualized in either the anatomic study or in the US examination. However, in CT images one caudal deep cervical LN (CDCLN) was found in 11 cats. This CDCLN lied in the fat

80 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

tissue that is slightly cranial to the thoracic inlet, between the trachea and the sternocephalicus muscle. The CDCLNs were identified as elongated, isoattenuating (72.7%) or heterogeneous (18.2%) structures, and commonly enhanced homogeneously after contrast administration.

Axillary lymph center (lymphocentrum axillare)

In the anatomic study, both right and left (AxLNs) were found in 3/6, and only the right AxLN was found in 1/6 cats, with a total of 7 AxLNs. The AxLNs were elongated or rounded (Fig. 5). The localization was immediately caudal to the axillary vessels at the level of the first intercostal space on each side of the thorax. In CT images, both AxLNs were identified in 28/30 cats, and only the right AxLN was seen in 1/30 cats, for a total of 57 AxLNs. The presence of a central hypoattenuating area with negative attenuation values (-29.7 HU) as the attenuation of fat tissue was identified. This hypoattenuating tissue within the AxLN was surrounded by an oval ring shaped tissue that corresponded to the normal attenuation of lymphatic tissue in the periphery. The regions of interests (ROIs) for the measurement of the Hounsfield units were placed, as possible, in this peripheral tissue. However, inclusion of part of the center in smaller LN made this challenging. This explains the negative HU in the average of attenuation for these LNs. The AxLNs were seen slightly hypoattenuating (R: 48.3%; L: 50%), heterogeneous (R: 27.6%; L: 25%) or isoattenuating (R: 17.2%; L: 17.9%) in pre-contrast images. After contrast administration, the right and the left AxLN presented homogeneous contrast enhancement in 41.4% and 46.4% of the cases respectively. Peripheral enhancement was present in 41.4% of the right and 35.7% of the left AxLNs. In US images, both AxLNs were identified in 29 cats, and only a right AxLN in 1 cat. The echogenicity of the AxLNs was the most variable among LNs. These LNs were more frequently isoechoic compared with the surrounding fat tissue. Additionally, a similar distribution of hyperechoic, hypoechoic, and heterogeneous echogenicity was also present in this group. The heterogeneous AxLNs presented a large hyperechoic center with a more

STUDIES 81

hypoechoic periphery (Figure. 5). A hyperechoic central line was identified in 13.6% of the AxLNs. Another component of this lymph center are the accessory axillary LNs (AAxLNs). Both right and left AAxLNs were found in 1/6 cats in the anatomic study, along the lateral thoracic vessels. On CT, the AAxLNs were easily identified. The number of these AAxLNs varied between sides (right and left) and among cats. At each side of the thorax 1 to 3 LNs were identified. A single LN located at the level of the third costocondral joint in the dorsal border of the pectoralis profundus muscle was seen in both sides in 21 cats. This LN appeared elongated or miscellaneous. Then one or two additional LNs were seen more caudally in 3 (2 right and 1 left), 4 (2 right and 2 left), 1 (3 right and 2 left), and 1 (2 right and 3 left) cats. The most caudal LN was seen almost reaching the costal arch and also in contact with the dorsal border of the pectoralis profundus muscle. The AAxLNs were more frequently isoattenuating, and less commonly slightly hypoattenuating to the surrounding muscles. After contrast, 100% of the AAxLN presented homogeneous contrast enhancement. The frequency of visualization of these LNs using US was low. A single LN was seen in both sides in 3 cats. One and two right AAxLNs were seen in 2 cats. The AAxLN appeared most frequently hypoechoic, followed by isoechoic or heterogeneous with less frequency. They were elongated or rounded with smooth borders. No hyperechoic central line was identified in any of the LNs (Figure. 6).

Dorsal thoracic lymph center (lymphocentrum thoracicum dorsale)

The aortic thoracic and the intercostal LNs are described as components of this lymph center. However, they were not visible in either the anatomy or in the imaging study.

Ventral thoracic lymph center (lymphocentrum thoracicum ventrale)

82 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

The only member of this lymph center that was identified was the sternal lymph node (SLN). The superficial cranial epigastric LN (former xiphoid LN) and the phrenic LN were not visible in the anatomy nor in the imaging study. The SLN was identified in 4/6 (66.7%) cadavers as an elongated or oval structure located at the dorsal aspect of the third sternebra and related with the internal thoracic vessels. The appearance of the SLNs on CT transverse images, and on multiplanar reconstruction, was similar to the description for the AxLNs. They were frequently ovoid with soft tissue attenuation peripherally and a hypoattenuating center. Regarding their attenuation, 71.4% were classified as heterogeneous; 14.3%, 9.5%, and 4.8% were hypoattenuating, slightly hypoattenuating, and isoattenuating, respectively. On postcontrast images, the LNs showed most frequently a peripheral enhancement, and less frequently heterogeneous contrast enhancement. The rest of the SLNs showed a slightly heterogeneous (27.3%) or homogeneous (4.5%) contrast enhancement. On the US images, the SLNs were identifed in 17/30 (56.7%) cats as isoechoic (47.0%), hypoechoic (35.3%), heterogeneous (11.8%) or hyperechoic (5.9%) with smooth borders. The SLN presented a hyperechoic central line with hypoechoic periphery in 17.6% of the LNs (Figure. 7).

Mediastinal lymph center (lymphocentrum mediastinale)

One cranial mediastinal LN (CrMLN) was identified in 5 cadavers in the anatomic study. A normal CrMLN was seen as a rounded or oval structure located in the mediastinum, between the trachea and the blood vessels. In CT images, one CrMLN was identified in 15 cats, and 2 CrMLNs were seen in 1 cat. These LNs were slightly hypoattenuating (66.7%) or isoattenuating (33.3%). After contrast administration, the CrMLNs presented a homogenous contrast enhancement (75%) or less frequently a slightly heterogeneous contrast enhancement (25%). Assessment of the CrMLN using US in healthy cats was not possible due to the impossibility to find an acoustic window through the normal pulmonary tissue.

STUDIES 83

Bronchial lymph center (lymphocentrum bronchale)

In the anatomic study, three tracheobronchial lymph nodes (TBLNs) were found corresponding to the right, left, and middle TBLNs, in each of the 6 cadavers. The right tracheobronchial LN (RTBLN) was visible between the main right bronchus and the azygos vein. The left tracheobronchial LN (LTBLN) was visible between the main left bronchus and the left pulmonary artery. The middle tracheobronchial LN (MTBLN) was found caudally to the carina. In CT images, the right, left, and middle TBLNs were observed in 6, 13, and 24 cats respectively. Combinations of the frequency of identification were as follows: the three TBLNs in 5 cats, the right and left TBLNs in 1 cat, the left and the middle in 5 cats, only the middle TBLN in 14 cats, and only the left TBLN in 2 cats. The TBLNs were not visible in 3 cats. The use of postcontrast images in the localization of these LNs was fundamental. Also, the MTBLN was assessed first in sagittal images obtained with multiplanar reconstruction. On these images, the identification of the LN between the carina and the pulmonary blood vessels was easier improving its identification in other planes; the same was true for the right and left TBLNs. In precontrast images, they were commonly isoattenuating and less commonly slightly hypoattenuating. Postcontrast images commonly showed homogeneous or less commonly slightly heterogeneous contrast enhancement (Figure. 8). Ultrasonographically it was impossible to obtain an acoustic window that allows the assessment of these lymph nodes in a healthy cat.

Statistical analysis

The means and SD for the length, height, and width measurements obtained in the anatomic and imaging studies of each lymph node are summarized in Table 1. Measurements data from TC, US, and anatomy showed differences when compared by the Wilcoxon signed rank test (TC vs. US) and by the Mann- Whitney U test (TC & US vs. Anatomy). Those differences are shown in Table 1.

84 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

The comparison of the calculated length vs. the measured length in MPR on CT images showed statistically significant differences for the RLMnLN, MTBLN and RTBLN; and for the right and left RLNs, the most dorsally located left DSCLN and the most cranial left AAxLN. The MPR length was chosen for further comparison with US and anatomic lengths because it was considered to produce a more understandable dimension of the LN and its relative position in the body than the calculated length (the length of LNs with relative oblique position could have been underestimated with the calculated length). The length of the LNs on CT was higher than US and anatomic lengths. The highest differences were found in the MRLNs, the length in CT was 6mm larger than the length in US and anatomy. A statistically significant difference was shown for the MnLN, MRLN, AxLN and SLNs between CT and US, and for the left MnLNs, MRLN, right and left VSCLN between CT and anatomy. In all the lymph nodes the lengths were higher on CT than on US or anatomy. The difference in lengths between US and anatomy were scarce (1 – 2mm) in most of the LN, but a statistically significant difference was observed only for the right MRLN. The length for this LN was higher in US than in anatomy (5mm). Ultrasonographically, the LNs were relatively wider than on CT and anatomy. The highest differences were presented for the MRLNs and SCLNs. A statistically significant difference was seen in most groups of LNs, except for PLNs, VSCLNs, and AAxLNs, when comparing CT with US. When CT widths were compared with anatomy, a statistically significant difference was found only for MnLN (apart from the right lateral MnLN), left MRLN, and right TBLN. The comparison of widths between US and anatomy showed statistically significant differences for the right MnLNs, MRLNs, the most ventral right and left DSCLN, and the AxLNs. The height was the most variable measurement among techniques showing the highest statistically significant differences among them. In the anatomic study, most of the LNs presented a height between 1.2 to 1.9mm, except the LMRLN (3.9mm) and SLN (2.3mm). These values were smaller compared with the height obtained with CT and US. The mean and SD of the Hounsfield units, and the description of the attenuation for each group of lymph nodes is reported in Table 2. Most of the

STUDIES 85

lymph nodes were isoattenuating or slightly hypoattenuating to the surrounding musculature, and showed homogeneous contrast enhancement. Only the SLN (71.4%) and the AxLNs (R: 27.6%; L: 25%) presented heterogeneous attenuation in pre-contrast and postcontrast images. The echogenicity and shapes percentages of the LNs are presented in Table 3. Most LNs were isoechoic or hypoechoic to surrounding tissue. The AxLN showed a high frequency of heterogeneous echogenicity compared with the rest of the lymph nodes. The majority of LNs were fusiform or rounded. Nevertheless, the MRLNs, DSCLNs, and AAxLNs had miscellaneous shape.

Discussion

In the anatomic study, the identification of most LNs was possible but not all of them could be properly dissected. A possible explanation for this is that some LNs can be very small and surrounded by a large amount of adipose tissue, making difficult to differentiate them from the fat (especially the superficial cervical, the deep cervical, and the dorsal thoracic lymph centers). The identification of the LNs from the head, neck, forelimb, and thoracic lymph centers in the cat was possible with diagnostic imaging techniques. In this study, CT showed a higher frequency of LNs identification in comparison to US and even to the anatomic study, and that frequency was improved in postcontrast studies. The advantage of the CT over the other techniques could be due to the fact that the LNs are well-vascularized structures. Organs with good blood supply have an optimal enhancement after the administration of contrast medium which makes them easier to depict, as previously reported for the TBLN (Dennler et al., 2013). However, when the LNs were located close to another well-vascularized structure (e.g. salivary gland), to differentiate it from the glandular tissue was challenging (e.g. PLN and LRLN). Another important fact is the body condition; adipose tissue provides a good contrast in CT images. When LNs are surrounded by a fair amount of fatty tissue, their visualization improves as previously reported in dogs (Beukers et al., 2013; Rossi, Patsikas, & Wisner, 2011). However, a large amount of fat around the LNs could reduce the differentiation during the US evaluation.

86 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

The PLNs were difficult to identify in all the techniques. Their close anatomic relationship with the parotid salivary gland most likely impaired the visualization and correct measurement of this LN in CT images; a similar situation has been reported in dogs (Kneissl & Probst, 2007). The identified PLNs were surrounded by an optimal amount of fatty tissue that clearly separated them from the parotid salivary gland. A facial hair trimming to assess the PLNs on US images was not performed for esthetical reasons. Therefore, the low resolution of the images did not allow an accurate assessment of these LNs. The mandibular LN could be easily visualized in CT and US in cats. Similar results have been reported in dogs (Kneissl & Probst, 2007; Nyman et al., 2005). The CT and US mean length of the right and left MnLNs in this study were shorter (10.86-11.35mm on CT and 9.01-9.87mm on US) than those reported by Sugimura et al. (1955) for the same LNs in cadavers (19mm- 24mm). However, the study of Sugimura et al. included very young animals (from 1 month to 6 year-old) and this could have an influence on their results. A previous report showed that young animals usually present bigger LNs in comparison with adults (Burns, Scrivani, Thompson, & Erb, 2008). Ultrasonographic features for the MnLNs in healthy cats have not been previously reported. In this study, an elongated, hypoechoic to isoechoic structure with smooth margins and a thin hyperechoic peripheral rim was identified. This description is similar to previous reports for dogs (Nyman et al., 2005). On CT images, the MnLNs were frequently seen with an isoattenuating appearance. The mean HU before and after contrast administration were similar to those reported in dogs (Kneissl & Probst, 2007). The features like attenuation, echogenicity, and size of the MRLNs in healthy cats using CT and US found in this study are similar to what have been previously described (Nemanic & Nelson, 2012; Oliveira, O’Brien, Matheson, & Carrera, 2012). The MRLNs measurements on CT showed the highest differences when compared with US and anatomy. The natural position of the MRLN makes that the transverse images obtained on CT do not correspond to an exact transverse slice of the lymph node but rather an oblique slice of it (Nemanic &

STUDIES 87

Nelson, 2012). Besides, in US the positioning of the probe to obtain appropriate images with a maximum long axis and short axis of the MRLN can be also slightly oblique. Nemanic & Nelson (2012) reported mild to moderate heterogeneous attenuation for the MRLN in healthy cats. In our study, slightly hypoattenuating to isoattenuating MRLN were more frequently found. To the authors’ knowledge, this is the first report with the description of normal US and CT features of the superficial and deep cervical lymph nodes. We found, two dorsal SCLN and one ventral SCLN related to the superficial cervical vessels with both imaging techniques (Saar & Getty, 1982). A differentiation was made between the two dorsal SCLNs that were found. The most dorsally located were smaller than the most ventrally located. On US assessment, the cranial border of the scapula and the fat tissue cranial to the supraspinatus muscle were used as an anatomic landmark to identify the DSCLN. This fat tissue is also visible on transverse CT images and helped in the identification of these LNs. However, the position of the forelimbs has an influence on the relative parallel or perpendicular disposition of the DSCLN with the spine. As described for the MRLN, this natural oblique position in transverse CT images could also explain the mean differences for length, width, and height among techniques for these LNs. The VSCLNs were identified more frequently with CT compared with the other techniques. The localization of these LNs on US images was challenging. The identification of the AxLNs and the AAxLNs was possible using CT and US following the anatomic landmarks obtained in the anatomic study, which was compatible with previous anatomic reports (Saar & Getty, 1982; Sugimura, Kudo, & Takahata, 1956; Tompkins, 1993). Sugimura et al. (1956) reported that the AxLNs were embedded in fatty tissue and were depressed, taking an ellipsoid shape. We assumed that the depression mentioned in that report might correspond in our study to the description of fatty hilus. This could explain the iso- to hyperechoic image in US and the hypoattenuating image in CT that made their visualization challenging (Nyman & O’Brien, 2007). There were a few differences in the size measurements of the AxLNs between anatomic and imaging studies. The height of the AxLNs was the measurement in which higher statistically significant differences were found

88 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

among techniques. The height in cadavers was significantly smaller than the height in US and CT images. Even though the cadavers were fresh and dissection was made within 24h of death, we consider that the amount of blood and lymph within the LNs were different and could influence these measurements. More studies are needed to establish this correlation in cats. The mean length of the right and left AxLNs and AAxLNs in this study is smaller than described by Sugimura (1956). Probably the difference could be explained again by the young age of the cats used in Sugimura’s study, where 16/24 cats were under 8 month-old. The ventral thoracic lymph center is formed by the sternal, phrenic, and cranial epigastric LNs (NAV, 2012; Saar & Getty, 1982; Sugimura, Kudo, & Takahata, 1959). Sugimura et al. (1959) reported a frequency of presence of 100% for the SLN, and only 4.5% and 27% for the phrenic and cranial epigastric LN, respectively. In our study, only the SLN was commonly identified. It is probable that the location of the phrenic LN, which is located close to the foramen venae cavae (Saar & Getty, 1982), makes it difficult to assess with US and could be very hard to delineate from the liver or the vein in CT images in healthy cats. However, its presence should be considered in cases of soft tissue nodules or masses around the foramen. The appearance of the normal SLN was similar to the descriptions for the AxLN (a hypoattenuating central area with peripheral soft tissue attenuation on CT image; a hyperechoic center on US images). In this study a normal SLN was 12.2 X 3.8 X 5.2mm (length X width X height) in CT and 7.3 X 4.9 X 3.7mm in US. These measurements were statistically different suggesting that CT measurements were larger compared with US. This result could be possibly influenced by the US scanning plane, displaying a relative oblique angle that was not a true sagittal plane of the LN; this effect has been previously described for the jejunal LNs in dogs (Agthe, Caine, Posch, & Herrtage, 2009). In a previous study of CT measurements of the SLN in 6 cats of 8 to 12 months-old, the results were surprisinly similar to those in our study 11.9 X 3.2 X 5.1 mm (Dennler et al., 2013). Factors that could contribute to these results are unclear. However, the cats included in Dennler’s study presented a relative adult weight, ranged 2.4 - 3.6 kg. In our study, the mean weight was 4.4 kg (SD 1.1).

STUDIES 89

During the assessment of the cranial mediastinum in the CT images, an ill- defined soft tissue attenuating structure was identified commonly surrounding the great vessels, trachea, and lymph nodes mixed with fatty tissue. The nature of this structure is unclear and histopathology of this region was not performed in any of the cats. We hypothesized that could be fibrous tissue, thymus gland residual tissue or mixture of lymphatic vessels, nerves, and mediastinal membranes. As a result, this soft tissue made difficult the delineation of the CrMLN. However, a slightly hypoattenuating, elongated node was identified in 53% of the normal cats. Normal CrMLNs were not visible in US due to the impossibility to find an acoustic window avoiding the pulmonary tissue. Besides, the small size of the LNs, and the amount of fat in the mediastinal space might also influence. The TBLNs showed a 100% frequency of identification in the anatomic study, however, identification in the imaging study was challenging. Endoscopic ultrasonography has been reported as a suitable procedure to assess the TBLNs in dogs (Gaschen, Kircher, & Lang, 2003; St-Vincent & Pharr, 1998). Unfortunately, in this study, endoscopic probes were not available to evaluate the TBLNs, therefore, no information about the US features was available. The CT images before and after the administration of contrast medium allowed the assessment of the TBLNs, being the middle TBLNs the one most frequently observed (80%). Sagittal multiplanar reconstruction of CT images was a very helpful tool to localize this LN and then it was easier to assess it in other planes. There are several limitations in this study. First, the number of cats in the anatomic study was low (n=6), this was due to the low number of cats that die of causes not related to either neoplastic or inflammatory processes. Second, an important variability in the identification of each group of lymph nodes per lymph center was present. It was not possible to ensure the same number of LNs in each cat and for both imaging techniques, which has a direct influence on the type of statistical test that can be applied. Third, all the healthy cats included in the imaging study were carefully evaluated in an attempt to avoid the inclusion of cats with lymphadenopathy. However, fine needle aspirates for histologic examination were not performed in order to avoid complications in these cats and for ethical reasons. Fourth, the assessment of the lymph

90 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

nodes features and size was made only one time by one of the authors (MT); therefore, the interobserver or intraobserver analysis could not be performed. In conclusion, the identification of lymph nodes in the head, neck, forelimb and thorax using US and CT is possible. Cats with a high body condition provide a good contrast in CT images to identify lymph nodes due to the amount of fat tissue around them. In very thin cats, the administration of contrast medium makes a lymph node recognizable because increases its delineation and differentiation from the surrounding muscles. The AxLN and the SLN present a relatively large and fatty hilus; this creates a different appearance on CT and US images compared to other lymph nodes. On CT images, a multiplanar reconstruction is a very useful tool that improves the accuracy of the assessment of the lymph nodes size avoiding the relatively oblique image in transverse slices of some lymph nodes (MRLNs, SCLNs) due to their natural position in the body. To the authors’ knowledge, this is the first report of the lymph nodes dimensions for the head, neck, forelimb, and thorax lymph centers using US and CT and comparing with an anatomic study in healthy cats

References

Agthe, P., Caine, A. R., Posch, B., & Herrtage, M. E. (2009). Ultrasonographic Appearance of Jejunal Lymph Nodes in Dogs Without Clinical Signs of Gastrointestinal Disease. Veterinary Radiology & Ultrasound, 50(2), 195–200. Beukers, M., Vilaplana Grosso, F., & Voorhout, G. (2013). Computed Tomographic Characteristics of Presumed Normal Canine Abdominal Lymph Nodes. Veterinary Radiology & Ultrasound, 54(6), 610-617. Burns, G. O., Scrivani, P. V, Thompson, M. S., & Erb, H. N. (2008). Relation Between Age, Body Weight, and Medial Retropharyngeal Lymph Node Size in Apparently Healthy Dogs. Veterinary Radiology & Ultrasound Ultrasound, 49(3), 277–281. Dennler, M., Bass, D. a., Gutierrez-Crespo, B., Schnyder, M., Guscetti, F., Di Cesare, A., Glaus, T. M. (2013). Thoracic computed tomography, angiographic computed tomography, and pathology findings in six cats experimentally infected with aelurostrongylus abstrusus. Veterinary Radiology and Ultrasound, 54(5), 459–469. Gaschen, L., Kircher, P., & Lang, J. (2003). Endoscopic ultrasound instrumentation, applications in humans, and potential veterinary applications. Veterinary Radiology and Ultrasound, 44(6), 665–680. Kneissl, S., & Probst, a. (2007). Comparison of computed tomographic images of normal cranial and upper cervical lymph nodes with corresponding E12 plastinated-embedded sections in the dog. Veterinary Journal, 174(2), 435–438. NAV. (2012). Nomina anatomica veterinaria. ( veterinary gross anatomical nomenclature International committee, Ed.) (Fifth Edit.). Oslo: ICVGAN. Nemanic, S., Hollars, K., Nelson, N. C., & Bobe, G. (2015). Combination of Computed Tomographic Imaging Characteristics of Medial Retropharyngeal Lymph Nodes and Nasal Passages Aids Discrimination Between Rhinitis and Neoplasia in Cats. Veterinary

STUDIES 91

Radiology & Ultrasound, 56(6), 617–627. Nemanic, S., & Nelson, N. C. (2012). Ultrasonography and noncontrast computed tomography of medial retropharyngeal lymph nodes in healthy cats. American Journal of Veterinary Research, 73(9), 1377–1385. Nemanic, S., & Nelson, N. C. (2012). Ultrasonography and noncontrast computed tomography of medial retropharyngeal lymph nodes in healthy cats. American Journal of Veterinary Research, 73(9), 1377–1385. Nyman, H. T., Kristensen, A. T., Skovgaard, I. M., & McEvoy, F. J. (2005). Characterization of normal and abnormal canine superficial lymph nodes using gray-scale B-mode, color flow mapping, power, and spectral doppler ultrasonography: A multivariate study. Veterinary Radiology and Ultrasound, 46(5), 404–410. Nyman, H. T., & O’Brien, R. T. (2007). The Sonographic Evaluation of Lymph Nodes. Clin Tech Small Anim Pract, 22(3), 128–137. Oliveira, C. R., O’Brien, R. T., Matheson, J. S., & Carrera, I. (2012). Computed tomographic features of feline nasopharyngeal polyps. Veterinary Radiology & Ultrasound. 53(4), 406–11. Rossi, F., Patsikas, M. N., & Wisner, E. R. (2011). Abdominal lymph nodes and lymphatic collecting system. In T. Schwarz & J. H. Saunders (Eds.),Veterinary computed tomography (pp. 371 – 379). Ames : Iowa: Wiley-Blackwell. Saar, L. I., & Getty, R. (1982). Sistema linfático de los carnívoros. In R. Getty (Ed.), S.Sisson - J.D.Grossman. Anatomía de los animales domésticos (Fifth edit., pp. 1811–1831). Barcelona, Spain : Masson. St-Vincent, R. S., & Pharr, J. W. (1998). Transesophageal ultrasonography of the normal canine mediastinum. Veterinary Radiology & Ultrasound, 39(3), 197–205. Sugimura, M., Kudo, N., & Takahata, K. (1955). Studies on the lymphonodi of cats: I. Macroscopical observations on the lymphonodi of heads and necks. Japanese Journal of Veterinary Research, 3(2), 90 – 104. Sugimura, M., Kudo, N., & Takahata, K. (1956). Studies of lymphonodi of cats: II. Macroscopical observations on the lymphonodi of the body surfaces, thoracic and pelvic limbs. Japanese Journal of Veterinary Research, 4(3), 101 – 112. Sugimura, M., Kudo, N., & Takahata, K. (1959). Studies on the lymphonodi of cats: IV. Macroscopical observations on the lymphonodi in the thoracic cavity and supplemental observations on those in the head and neck. Japanese Journal of Veterinary Research, 7(1-4), 27 – 51. Tompkins, M. B. (1993). Lymphoid system. In Atlas of feline anatomy for veterinarians (pp. 113 – 126). Philadelphia [etc.] : W.B. Saunders Company.

Table 1. Mean and SD for length (MPR, Calculated), width, and height of the LNs of the head, neck, forelimb, and thorax. Comparison among techniques. Length (mm)

CT- CT- MPR US Lymph CT-MPR CT-Calc. US Anatomy MPR MPR vs. vs. node m (SD) m (SD) m (SD) m (SD) vs. vs. Calc. (A) Anatomy (B) US (A) Anatomy (B) RP 5.56 (1.26) 6.57 (1.85) - 5.00 (0.85) LP 5.62 (2.05) 6.37 (2.17) 10.00 (NC) 4.90 (0.26) RMMn 11.04 (3.03) 11.41 (3.47) 9.83 (2.84) 8.28 (4.82) * ** RLMn 10.86 (3.23) 11.63 (3.92) 9.71 (2.28) 9.87 (2.24) * LMMn 11.35 (2.77) 11.42 (3.65) 9.01 (2.67) 10.33 (3.28) ** ** LLMn 11.32 (2.45) 11.38 (3.35) 9.87 (2.08) 9.67 (2.75) ** ** RMR 21.25 (4.27) 18.63 (6.19) 13.38 (3.84) 7.63 (4.65) ** ** ** * LMR 20.42 (3.38) 17.77 (5.85) 14.26 (3.82) 13.90 (4.95) ** ** * RDSC1 8.18 (3.65) 8.35 (3.57) 7.37 (4.39) - RDSC2 6.73 (3.19) 8.30 (4.22) 8.38 (3.60) 9.35 (4.11) RVSC 7.92 (2.21) 8.18 (2.70) 10.15 (1.20) 11.25 (2.88) * LDSC1 8.44 (4.02) 8.04 (3.98) 6.34 (2.39) - LDSC2 6.03 (3.00) 7.54 (3.15) 7.48 (3.55) 7.55 (3.18) ** LVSC 8.28 (2.07) 8.72 (2.40) 9.70 (1.27) 12.07 (1.54) ** CDC 7.49 (3.81) 8.13 (3.64) - - RAx 8.82 (1.98) 9.62 (4.01) 7.80 (1.96) 8.28 (3.80) ** LAx 9.24 (2.01) 9.73 (2.47) 7.23 (1.73) 8.80 (2.79) ** RAAx1 14.18 (4.65) 14.48 (4.62) 9.30 (5.63) 17.00 (NC) RAAx2 6.69 (2.47) 7.80 (3.92) 6.30 (NC) - LAAx1 13.43 (6.07) 13.71 (5.57) 9.77 (2.65) 5.9 (NC) ** LAAx2 9.54 (5.28) 9.80 (5.14) - - LAAx3 4.60 (NC) 5.00 (NC) - - S 12.24 (3.25) 13.59 (5.43) 7.35 (2.42) 8.62 (3.83) ** CrM 8.14 (4.03) 8.02 (4.09) - 5.66 (2.44) MTB 4.82 (1.26) 5.59 (1.88) - 6.32 (3.07) * RTB 4.03 (0.77) 4.40 (1.13) - 5.20 (2.41) * LTB 5.62 (2.12) 5.86 (2.43) - 6.45 (3.15)

Continuation table 1. Mean and SD for length (MPR, Calculated), width, and height of the LNs of the head, neck, forelimb, and thorax. Comparison among techniques.

Width (mm) Height (mm)

CT CT US CT CT US Lymph CT US Anatomy CT US Anatomy vs vs vs vs vs vs node m (SD) m (SD) m (SD) m (SD) m (SD) m (SD) US (A) Anatomy (B) Anatomy (B) US (A) Anatomy (B) Anatomy (B) RP 3.52 (0.63) - 3.17 (1.36) 4.43 (1.24) - 1.43 (0.51) * LP 3.19 (0.68) 9.30 (NC) 2.23 (0.55) 4.74 (1.15) 3.80 (NC) 1.30 (0.61) * RMMn 5.71 (1.78) 6.83 (2.07) 4.38 (1.46) ** ** 2.87 (1.13) 2.66 (0.95) 1.60 (0.30) ** ** RLMn 6.53 (1.65) 7.85 (1.95) 4.63 (0.82) ** ** ** 3.22 (0.95) 2.84 (0.81) 1.72 (0.41) * ** ** LMMn 5.40 (1.67) 6.23 (1.53) 4.42 (1.26) ** ** 2.84 (0.99) 2.51 (0.55) 1.70 (0.24) * ** LLMn 6.88 (1.82) 7.60 (1.93) 4.23 (0.90) ** ** 3.43 (0.98) 3.05 (0.76) 1.52 (0.38) ** ** RMR 6.10 (8.54) 13.33 (3.22) 4.67 (3.76) ** * 11.92 (2.41) 4.49 (1.14) 1.80 (0.82) ** ** ** LMR 4.79 (2.03) 11.62 (2.46) 7.38 (2.73) ** * ** 11.95 (2.47) 4.40 (1.37) 3.94 (3.16) ** ** RDSC1 4.56 (1.94) 8.21 (3.12) - ** 4.85 (2.96) 2.51 (0.70) - ** RDSC2 4.55 (1.05) 10.38 (5.21) 2.75 (2.09) ** ** 11.53 (4.80) 2.88 (0.79) 1.32 (0.85) ** ** * RVSC 3.72 (0.83) 6.60 (2.12) 3.12 (1.00) 4.00 (1.96) 3.40 (0.57) 1.20 (0.44) ** LDSC1 4.73 (1.96) 7.50 (3.97) - ** 5.40 (2.26) 2.34 (0.66) - ** LDSC2 4.36 (1.06) 7.98 (4.47) 2.85 (1.70) * ** 10.36 (4.24) 2.40 (0.64) 1.25 (0.76) * ** * LVSC 3.86 (1.21) 8.60 (3.11) 3.6 (1.79) 4.04 (1.70) 4.60 (0.28) 1.45 (0.98) ** CDC 3.19 (1.11) - - 3.38 (2.43) - - RAx 3.18 (1.02) 4.93 (1.08) 3.25 (1.21) ** * 4.03 (1.48) 3.30 (0.97) 1.20 (0.36) * ** ** LAx 3.91 (1.45) 4.73 (1.20) 2.40 (0.66) * ** 4.51 (1.65) 3.52 (1.12) 1.20 (0.17) * ** ** RAAx1 3.03 (1.04) 5.88 (2.14) 1.40 (NC) 4.38 (1.98) 2.48 (1.31) 0.70 (NC) RAAx2 2.81 (1.28) 6.50 (NC) - 3.50 (1.24) 1.30 (NC) - LAAx1 3.09 (1.13) 6.50 (1.59) 3.20 (NC) 3.98 (1.73) 2.87 (0.85) 2.00 (NC) LAAx2 2.85 (1.00) - - 3.65 (0.59) - - LAAx3 2.70 (NC) - - 4.10 (NC) - - S 3.88 (1.27) 4.96 (1.21) 4.28 (2.13) 5.20 (1.46) 3.72 (1.18) 2.25 (0.42) * ** * CrM 3.26 (1.24) - 3.84 (2.13) 3.42 (1.67) - 1.88 (1.25) * MTB 3.23 (0.99) - 2.55 (1.13) 2.42 (0.59) - 1.23 (0.50) ** RTB 3.87 (1.19) - 2.03 (0.61) ** 3.25 (0.57) - 1.20 (0.51) ** LTB 3.19 (1.12) - 2.60 (1.26) 3.01 (1.00) - 1.15 (0.33) **

RP: right parotid, LP: left parotid, RMMn: right medial mandibular, RLMn: right lateral mandibular, LMMn: left medial mandibular, LLMn: left lateral mandibular, RMR: right medial retropharyngeal, LMR: left medial retropharyngeal, RDSC1 and RDSC2: right dorsal superficial cervical, RVSC: right ventral superficial cervical, LDSC1 and LDSC2: left dorsal superficial cervical, LVSC: left ventral superficial cervical, CDC: caudal deep cervical, RAx: right axillary, LAx: left axillary, RAAx1 and RAAx2: right accessory axillary, LAAx1, LAAx2 and LAAx3: left accessory axillary, S: sternal, CrM: cranial mediastinal, MTB: medial tracheobronchial, RTB: right tracheobronchial, LTB: left tracheobronchial. NC: not calculated. * p-value < 0.05 ** p-value < 0.01 (A) Wilcoxon Signed Rank Test (B) Mann-Whitney U Test

Table 2. Computed tomographic characteristics of the lymph nodes of the head, neck, forelimb, and thorax of healthy cats. Lymph HU Precontrast HU Postcontrast Attenuation Precontrast (%) Attenuation Postcontrast (%)

node Mean SD Min Max Mean SD Min Max Iso S Hypo Hypo Hyper Heter Hom S Het Het Peri

RP 36.52 9.53 18.67 50.67 131.67 55.46 65.00 257.67 88.89 11.11 0.00 0.00 0.00 100.00 0.00 0.00 0.00 LP 33.94 7.16 20.33 44.00 100.03 30.99 45.33 154.33 83.33 16.67 0.00 0.00 0.00 83.33 0.00 16.67 0.00 RMMn 38.74 11.98 12.33 60.67 137.07 33.46 60.00 209.67 63.33 20.00 16.67 0.00 0.00 90.00 6.67 3.33 0.00 RLMn 36.38 12.48 12.67 63.33 130.84 34.81 52.00 203.67 63.33 20.00 16.67 0.00 0.00 86.67 10.00 3.33 0.00 LMMn 38.54 14.52 6.00 57.67 134.43 33.64 68.33 195.67 63.33 20.00 16.67 0.00 0.00 86.67 6.67 6.67 0.00 LLMn 37.08 10.86 12.00 54.33 130.15 31.64 65.67 182.67 60.00 23.33 16.67 0.00 0.00 80.00 13.33 6.67 0.00 RMR 44.50 6.63 25.67 61.67 133.14 25.42 88.33 183.00 26.67 60.00 13.33 0.00 0.00 40.00 53.33 6.67 0.00 LMR 43.32 7.46 24.67 55.67 131.99 27.61 90.00 184.33 26.67 60.00 13.33 0.00 0.00 37.93 55.17 6.90 0.00 RDSC1 22.77 20.46 -39.67 50.00 92.71 30.12 30.33 159.67 52.00 28.00 8.00 0.00 12.00 88.00 0.00 0.00 12.00 RDSC2 31.71 15.52 -11.67 60.67 115.79 20.31 71.67 154.67 56.67 36.67 3.33 0.00 3.33 90.00 3.33 6.67 0.00 RVSC 32.76 13.84 1.00 59.00 122.14 28.99 60.00 177.67 44.83 37.93 10.34 0.00 6.90 79.31 10.34 6.90 3.45 LDSC1 25.67 23.48 -25.33 62.33 95.01 35.81 5.33 161.67 54.17 29.17 4.17 0.00 12.50 87.50 0.00 0.00 12.50 LDSC2 31.90 14.06 -9.00 58.33 116.09 24.08 67.00 154.67 64.29 32.14 3.57 0.00 0.00 96.43 0.00 3.57 0.00 LVSC 35.09 15.25 -19.67 51.33 119.36 31.78 52.00 176.00 56.00 28.00 4.00 0.00 12.00 80.00 8.00 8.00 4.00 CDC 33.03 19.80 -13.33 54.00 94.64 23.99 42.67 128.33 72.73 9.09 0.00 0.00 18.18 63.64 18.18 9.09 9.09 RAx 15.14 20.37 -29.67 54.33 77.62 36.51 17.33 147.67 17.24 48.28 3.45 3.45 27.59 41.38 17.24 0.00 41.38 LAx 20.08 19.43 -19.33 53.67 80.88 33.38 32.00 137.00 17.86 50.00 3.57 3.57 25.00 46.43 17.86 0.00 35.71 RAAx1 31.43 17.03 -16.67 56.67 97.66 25.92 29.67 143.00 70.00 30.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 RAAx2 20.67 23.36 -19.33 51.00 91.00 31.96 38.33 134.33 77.78 22.22 0.00 0.00 0.00 100.00 0.00 0.00 0.00 LAAx1 30.09 19.75 -35.67 57.00 96.44 26.00 26.67 150.67 70.00 30.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 LAAx2 31.29 15.78 -4.00 48.67 99.75 34.99 50.00 154.33 87.50 12.50 0.00 0.00 0.00 100.00 0.00 0.00 0.00 LAAx3 40.33 NC 40.33 40.33 86.33 NC 86.33 86.33 100.00 0.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 S -9.43 27.75 -55.67 33.33 82.23 45.28 -12.33 181.67 4.76 14.29 9.52 0.00 71.43 4.55 27.27 0.00 68.18 CrM 30.21 11.92 10.67 52.00 96.08 37.40 33.00 174.67 75.00 25.00 0.00 0.00 0.00 75.00 25.00 0.00 0.00 MTB 20.10 18.93 -26.00 50.67 85.76 26.59 43.33 134.67 62.50 33.33 4.17 0.00 0.00 87.50 12.50 0.00 0.00 RTB 28.95 15.92 12.33 56.67 105.66 13.04 84.33 120.00 66.67 33.33 0.00 0.00 0.00 83.33 16.67 0.00 0.00 LTB 18.53 22.35 -39.67 39.67 90.25 47.31 4.67 178.67 58.33 33.33 8.33 0.00 0.00 91.67 8.33 0.00 0.00 RP: right parotid, LP: left parotid, RMMn: right medial mandibular, RLMn: right lateral mandibular, LMMn: left medial mandibular, LLMn: left lateral mandibular, RMR: right medial retropharyngeal, LMR: left medial retropharyngeal, RDSC1 and RDSC2: right dorsal superficial cervical, RVSC: right ventral superficial cervical, LDSC1 and LDSC2: left dorsal superficial cervical, LVSC: left ventral superficial cervical, CDC: caudal deep cervical, RAx: right axillary, LAx: left axillary, RAAx1 and RAAx2: right accessory axillary, LAAx1, LAAx2 and LAAx3: left accessory axillary, S: sternal, CrM: cranial mediastinal, MTB: medial tracheobronchial, RTB: right tracheobronchial, LTB: left tracheobronchial. Iso: isoattenuating, S Hypo: slightly hypoattenuating, Hypo: hypoattenuating, Hyper: hyperattenuating. Hom: homogeneous, S Het: slightly heterogeneous, Het: heterogeneous, Peri: peripheral enhancement. NC: not calculated.

Table 3. Ultrasonographic features of the lymph nodes of the head, neck, forelimb, and thorax in healthy cats.

Lymph Echogenicity (%) Shape (%)

node Isoechoic Hypoechoic Hyperechoic Heterogeneous Rounded Elongated Miscellaneous

RP* ------LP 0.00 100.00 0.00 0.00 0.00 100.00 0.00 RMMn 10.00 83.33 0.00 6.67 3.33 96.67 0.00 RLMn 10.00 90.00 0.00 0.00 0.00 100.00 0.00 LMMn 10.00 86.67 0.00 3.33 0.00 100.00 0.00 LLMn 6.67 93.33 0.00 0.00 0.00 100.00 0.00 RMR 23.33 66.67 0.00 10.00 0.00 83.33 16.67 LMR 13.33 73.33 0.00 13.33 0.00 86.67 13.33 RDSC1 3.45 82.76 0.00 13.79 6.90 89.66 3.45 RDSC2 11.11 88.89 0.00 0.00 0.00 88.89 11.11 RVSC 0.00 50.00 0.00 50.00 50.00 50.00 0.00 LDSC1 3.85 76.92 3.85 15.38 11.54 80.77 7.69 LDSC2 16.67 66.67 0.00 16.67 0.00 83.33 16.67 LVSC 5000 0.00 0.00 50.00 100.00 0.00 0.00 CDC* ------RAx 46.67 16.67 16.67 20.00 63.33 36.67 0.00 LAx 37.93 17.24 17.24 27.59 65.52 34.48 0.00 RAAx1 20.00 60.00 0.00 20.00 0.00 80.00 20.00 RAAx2 0.00 100.00 0.00 0.00 0.00 100.00 0.00 LAAx1 33.33 66.67 0.00 0.00 0.00 100.00 0.00 LAAx2* ------LAAx3* ------S 47.06 35.29 5.88 11.76 52.94 47.06 0.00 CrM* ------MTB* ------RTB* ------LTB* ------* LNs that were not identified in US. RP: right parotid, LP: left parotid, RMMn: right medial mandibular, RLMn: right lateral mandibular, LMMn: left medial mandibular, LLMn: left lateral mandibular, RMR: right medial retropharyngeal, LMR: left medial retropharyngeal, RDSC1 and RDSC2: right dorsal superficial cervical, RVSC: right ventral superficial cervical, LDSC1 and LDSC2: left dorsal superficial cervical, LVSC: left ventral superficial cervical, CDC: caudal deep cervical, RAx: right axillary, LAx: left axillary, RAAx1 and RAAx2: right accessory axillary, LAAx1, LAAx2 and LAAx3: left accessory axillary, S: sternal, CrM: cranial mediastinal MTB: medial tracheobronchial, RTB: right tracheobronchial, LTB: left tracheobronchial.

STUDIES 97

Figure1. Parotid lymph node. a. Image of the dissection showing the localization of the parotid LN (arrow) rostral to the parotid salivary gland (asterisk) and superficial to the masseter muscle (M). b. Ultrasonographic image showing an elongated, hypoechoic parotid LN between cursors, the masseter muscle (M) is seen in the far field. c – d. CT images indicating the localization of an isoattenuating parotid lymph node (arrow) in the precontrast image (c) with homogeneous contrast enhancement pattern (d). The masseter muscle (M) and the zygomatic arch (asterisk) are indicated.

98 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

Figure 2. Mandibular lymph nodes. a. Image of the dissection showing the localization of the mandibular LNs (long arrow = lateral; arrow head = medial) rostral to the mandibular salivary gland (S). The linguofacial (F) and maxillary (MV) veins are indicated. b. Ultrasonographic image showing a transverse plane of a right lateral mandibular LN between cursors. Color Doppler shows the linguofacial vein (blue) medial to the LN. c – d. CT images indicating the localization of an isoattenuating mandibular LNs (long arrow= left lateral; arrow head= left medial) in the precontrast image (c) and homogeneous enhancement in the postcontrast image (d). The linguofacial vein is indicated (asterisk). The masseter muscle (M), digastric muscle (Dg), and tympanic bulla (B) are indicated.

STUDIES 99

Figure 3. Retropharyngeal lymph nodes. a. Image of the dissection showing the localization of the medial retropharyngeal LNs (arrow) ventral to C1 (asterisk). The carotid artery (CA), and the longus colli muscle (LC) are indicated. b. Ultrasonographic image showing an elongated and hypoechoic medial retropharyngeal LN (between cursors) located caudal to the mandibular salivary gland (S), ventro-medial to the partially seen carotid artery (CA) and the longus colli muscle (LC). The sternocephalicus muscle (SC) is indicated. c – d. CT images indicating the localization of the medial retropharyngeal LN (arrow); in the precontrast image (c) the elongated isoattenuating node is visible caudo-ventral to the tympanic bulla (B), ventral to the longus colli muscle (LC) at the level of C1 (asterisk). The postcontrast image (d) shows a slightly heterogeneous contrast enhancement pattern. The sternocephalicus muscle (SC) is indicated.

100 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

Figure 4. Superficial cervical lymph node. a. Image of the dissection showing the localization of the superficial cervical LNs (long arrow = dorsal; short arrow = ventral), cranial to the scapula (delineated area). The superficial cervical vessels (asterisk) are seen between the two nodes. The brachiocephalic (BC) and supraspinatus (SS) muscles and the jugular vein (JV) are indicated. b. Ultrasonographic image showing a fusiform, hypoechoic dorsal superficial cervical LN between cursors deep to the omotransversarius muscle (OT). Medial to the LN, the splenius muscle (Sp) is indicated. c – d. CT images indicating the localization of a slightly hypoattenuating dorsal superficial cervical LN (arrow) in the precontrast image (c), with a homogeneous contrast enhancement pattern in the postcontrast image (d). The sixth cervical vertebra (C6), the jugular vein (asterisk), the brachiocephalic (BC), omotransversarius (OT) and splenius (Sp) muscles are indicated.

STUDIES 101

Figure 5. Axillary lymph nodes. a. Image of the dissection showing the localization of the axillary LN (arrow) embedded in the fat caudally to the axillary vessels (axillary vein= AV; axillary artery= AA). b. Ultrasonographic image showing a heterogeneous axillary LN (between cursors). A large central hyperechoic center with a hypoechoic periphery is seen. The LN is located caudal to the axillary vessels (axillary vein= AV; axillary artery= AA), the pectoralis muscle (P) is ventral to the LN (top part of the image is ventral. Part of the brachial plexus (delineated area) and the scalenus muscle (S) are indicated. c – d. CT images indicating the localization of a slightly hypoattenuating axillary LN (arrow) in the precontrast image (c) with a homogeneous contrast enhancement pattern in the postcontrast image (d). The first thoracic vertebra (T1), the pectoralis (P) and scalenus (S) muscles are indicated.

102 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

Figure 6. Accessory axillary lymph node. a. Image of the dissection showing the localization of the accessory axillary LN (arrow) around the 4th intercostal space, along the lateral thoracic vessels (asterisk), between the pectoralis (P) and serratus ventralis thoracicus (SVT) muscles. b. Ultrasonographic image showing an elongated, hypoechoic left accessory axillary LN between cursors, deep to the cutaneous muscles (asterisks). The third, fourth, and fifth ribs (R3, R4, R5) and the lung field (L) are indicated. c – d. CT images indicating the localization of a slightly hypoattenuating accessory axillary LN (arrow) in precontrast (c) with a homogeneous contrast enhancement pattern in the postcontrast image (d), dorsal to the pectoralis (P) muscle and between the serratus ventralis thoracicus (SVT) and the latissimus dorsi (asterisk).

STUDIES 103

Figure 7. Sternal lymph node. a. Image of the dissection after parasternal thoracotomy showing the localization of the sternal LN (long arrow) along the internal thoracic vessels (short arrow) at the cranioventral aspect of the thorax. The sternum (St) has been ventrally pulled to allow the visualization of the SLN. b. Ultrasonographic image showing a rounded LN with a hyperechoic center and hypoechoic periphery between cursors. The second and third ribs (asterisks) and the pectoralis muscle (P) are indicated. c – d. CT sagittal images indicating the localization of the sternal LN in pre (c) and postcontrast (d) images (long arrows). The hypoattenuating center compatible with a fatty hilus is clearly visible. The third sternebra (asterisk) and internal thoracic vessels (short arrow) are indicated.

104 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb

Figure 8. Tracheobronchial lymph nodes. a. Image of the dissection showing the localization of the left (1), middle (2), and right (3) tracheobronchial LNs in relation to the carina (asterisk) and the main bronchi (short arrow= left; long arrow= right). The aorta (arrow head), and the left (LL) and right (RL) lungs are indicated. b – d. CT images (b, transverse; c – d, sagittal) indicating the localization of an isoattenuating middle tracheobronchial LN (arrow) in the precontrast image (c) with a homogeneous contrast enhancement pattern in the postcontrast images (b & d), caudal to the carina (C) and ventral to the esophagus (E, with moderate amount of gas in the lumen). The heart (H) and sixth thoracic vertebra (Th6) are indicated. In b, the main bronchi (red asterisk= right; blue asterisk= left), the main pulmonary arteries (yellow asterisk= right; green asterisk= left), the aorta (Ao), and the right (RL) and left LL) lungs are indicated.

4.2. Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb.

107 STUDIES

Abstract

The purposes of the study were to compare the dimensions of the abdomen and hindlimb lymph nodes measured with ultrasonography (US) and computed tomography (CT) with measurements obtained from an anatomic study, and to describe the features of these LNs using US and CT. In the anatomic study, 6 cadavers were dissected and in the imaging study 30 healthy cats were prospectively enrolled. In the dissection of the cadavers, at least one lymph node per lymph center was identified. Only the caudal epigastric LN (from the inguinofemoral lymph center) and the sacral LN (from the iliosacral lymph center) were not identified in any of the cadavers. All the lymph centers were visualized using CT with a variable frequency of identification. Factors like the amount of adipose tissue and the use of contrast medium subjectively improved the visualization of the lymph nodes. However, these factors were not included in the statistical analysis. Ultrasonographically, it was possible to identify almost all the LNs from each abdominal and hindlimb lymph center. The lumbar aortic, the internal iliac, the caudal epigastric, and the ischiatic LNs were not identified using US. Measurements of the length, width, and height were performed in each technique. Commonly, the measurements with CT were larger when compared with US and anatomy, although the main statistical differences were present in the comparison of height among techniques. The shape of the LNs was similar among techniques. Most of the identified LNs were elongated and a rounded shape was most common in hepatic, splenic, pancreaticoduodenal, colic, and popliteal LNs. A miscellaneous shape was commonly present in the jejunal LNs. The appearance of the LNs was mainly homogeneous in CT (before and after contrast administration) and in US. However, some LNs presented a more hyperattenuating periphery with a hypoattenuating center (before and after contrast administration). Something similar was detected with US, were some LNs presented a more hypoechoic / isoechoic periphery with a hyperechoic center (when compared with the surrounding tissue). Findings in this study indicated that the assessment of the LNs of the abdomen and hindlimb in the cat is possible with CT and US. The measurements and features reported are proposed as reference values.

108 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Introduction

The abdominal lymph centers in the cat are divided into parietal and visceral groups as in dogs (Beukers, Vilaplana Grosso, & Voorhout, 2013; Bezuidenhout, 2013). The parietal group has four lymph centers (lumbar, iliosacral, inguinofemoral, and ischiatic), and the visceral group has three lymph centers (celiac, cranial and caudal mesenteric) (Saar & Getty, 1982). Previous studies regarding the ultrasonographic assessment of the abdominal cavity in the cat reported that the most frequently identified lymph nodes from the visceral group are the gastric, hepatic, pancreaticoduodenal, jejunal, ileocecal, and colic lymph nodes; and from the parietal group are the medial iliac and the superficial . The lymph nodes less frequently identified in the visceral group are the splenic and caudal mesenteric lymph nodes; and in the parietal group are the lumbar aortic, renal, internal iliac (formally called hypogastric Nomina Anatomica Veterinaria (NAV), (2012)), sacral, and caudal epigastric lymph nodes (D’Anjou, 2008; Schreurs et al., 2008). The peripheral lymph nodes of the abdomen are described as fusiform and slender in shape with ultrasonography (US). When compared with surrounding fat tissue, they appeared slightly round in shape, with regular margins and hypoechoic. Deep abdominal lymph nodes are more rounded and elongated in shape, slightly hypoechoic to surrounding peritoneum and fat and also with regular margins (D’Anjou, 2008; Mattoon, Berry, & Nyland, 2015; Nyman & O’Brien, 2007; Schreurs et al., 2008; Widmer, Mattoon, & Nyland, 2015). Getty et al. (1982) reported two lymph centers for the hindlimb in the cat, the iliofemoral and the popliteal lymph centers. The popliteal lymph node is described as an oval or rounded shaped node of variable size (short axis range of 2.8 – 6.5mm; long axis range of 4.3 – 12.0mm) in ultrasound (Lee et al., 2012).

There is scarce literature available on normal computed tomographic appearance of the abdominal lymph nodes in the cat. In dogs, abdominal lymph nodes have been described as homogeneously attenuating structures commonly elongated in shape. Some lymph nodes were slightly irregular or

109 STUDIES relatively more hyperattenuating in the periphery than centrally before and after contrast administration (Beukers et al., 2013).

To the author’s knowledge, studies comparing the dimensions and features of abdominal and hindlimb LNs between the US and CT techniques are lacking.

The aims of this study were: (i) to compare the size of the abdomen and hindlimb LNs obtained with US and CT in a group of healthy adult cats with measurements obtained from an anatomic study; (ii) to describe the features of the LNs in the abdomen and hindlimb using US and CT in a group of healthy adult cats.

Materials and methods

The ethical committee of the Universitat Autònoma de Barcelona approved this study; reference number CEAAH 2255 of September 2013. The owner consent from all the patients and cadavers included were also obtained.

Anatomical study

Animals

The same six feline cadavers included in the part I of the study (Tobón Restrepo et al., 2016) were used to perform the dissection of the abdominal and hindlimb lymph centers. The cats were prospectively included during January 2013 to June 2015 from the cadavers referred to the pathology department of the Universitat Autònoma de Barcelona within 24 hours of death.

Continuing the dissection, the skin was removed for the evaluation of the abdominal cavity as described in part I (Tobón Restrepo et al., 2016). A small incision was made immediately caudal to the xiphoid process in the linea alba with a 24 scalpel blade. The incision was continued caudally with Mayo

110 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

scissors until the pubic bone. Afterwards, the abdominal muscles were cut following the costal arch until the spine. Then, blunt dissection of the abdominal lymph nodes was performed with special care in avoiding great vessels rupture. Then, the skin was removed from the hindlimbs with an incision in the cranial aspect of the thigh until the talus. After that, the incision surrounded the talus and the skin was pulled off. The lymph nodes were searched following the previous anatomic descriptions (Saar & Getty, 1982; Schreurs et al., 2008; Tompkins, 1993). Measurements were recorded with the same methods as described in part I (Tobón Restrepo et al., 2016).

Imaging study

Animals

In this study, the same group of healthy cats described in the part I was used. Healthy cats older than 1 year of age were recruited at the Fundació Hospital Clinic Veterinari of the Universitat Autònoma de Barcelona (FHCV-UAB) from staff, students and owners. The healthy state of the cats was determined following the same test and protocols described in part I (Tobón Restrepo et al., 2016).

Computed tomography and ultrasonography

Preparation, anesthetic protocol, CT and US settings were the same used and described in part I (Tobón Restrepo et al., 2016).

In this study, the assessment of the abdominal cavity to identify the lymph nodes was performed following the anatomic references and literature. Using multiplanar reconstruction (MPR) on CT images, the lymph nodes along the intestinal tract (e.g. colic LNs), or along the great vessels of the abdomen (e.g. medial iliac LNs), were numbered from orad to aborad, and from cranial to caudal, respectively. In the case of the ileocecal LNs, the most ventral was

111 STUDIES numbered as 1 and the dorsal as 2. In the case of the jejunal LNs, the LN at the medial or dorsal aspect of the jejunal vessels was numbered as 1 and the one lateral or ventral as 2. All these indications were done also during the US examination and in the anatomic study in order to get an accurate comparison of the different lymph centers.

Image analysis of the CT and the US characteristics and measurements from each lymph node identified were recorded following the same criteria reported in part I (Tobón Restrepo et al., 2016).

Statistical analysis

Microsoft Excel (2010)12 was used to digitalize the data. Statistical analyses were performed using the free available statistics software R (2015)13. The frequency of LNs identification, mean and SD of attenuation values pre- and postcontrast administration, echogenicity, and the mean and SD of LNs measurements was calculated. Wilcoxon Signed Rank Test was used to compare the pair distribution between the calculated length and the length obtained in the multiplanar reconstruction (length MPR) of the LNs on CT images. After this, the length MPR was used in the pair comparison with the US. The rest of the LNs measurements (width and height) between TC and US were also compared with Wilcoxon Signed Rank Test. Mann-Whitney U test was used to compare the pair distribution of the LN measurements (length MPR, width, and height) between TC and anatomy, and between US and anatomy. Each measurement was compared individually for each lymph center and not for the whole sample of identified LNs (No Post-Hoc corrections were used). A P value <0.05 was considered statistically significant.

12 Microsoft office Excel, 2010 13 R versión 3.2.3 (2015-12-10). Copyright © 2015, the R foundation for statistical computing.

112 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Results

Animal description

Anatomic study: six feline cadavers were included. Causes of death were not related to neoplastic or inflammatory diseases according to the necropsy (heart failure (n=2), kidney failure (n=2), poisoning (n=1) and unknown (n=1)). Average age was 6.8 years (range 1 – 16). Five cats were domestic shorthairs and one cat was a British longhair. Imaging study: thirty cats were recruited. Age and weight averages were 3.7 years (range 1.5 – 17) and 4.4kg (SD1.1) respectively. Twenty-nine cats were domestic shorthairs and 1 cat was a Persian. The group included 16.7% entire males (n=5), 20.0% neutered males (n=6), 30.0% entire females (n=9), and 33.3% neutered females (n=10). Biochemical determinations and complete blood count were within normal limits. All cats were negative for FIV/FeLV and Bartonella sp. tests.

Table 1 shows the mean and SD for the length, width, and height, as well as the results of the statistical comparisons among techniques; table 2 shows the attenuation values (Hounsfield units), and table 3 the ultrasonographic features.

Celiac lymph center (lymphocentrum celiacum)

Gastric lymph nodes (GLNs): in the anatomic study, one round or ovoid GLN was found in 5 cadavers located always in the omentum of the gastric lesser curvature. A fair amount of fat tissue was covering the GLNs in most of the cases.

The frequency of identification of the GLNs on the CT images was higher than US. One GLN was identified in 22 cats, and two GLNs were identified in 6 cats. The GLN was not identified in 2 cats. The appearance of these LNs in

113 STUDIES precontrast images was more frequently isoattenuating or slight hypoattenuating, and less frequently hypoattenuating or heterogeneous. A homogeneous contrast enhancement was most frequently seen, but in 14.29% of the LNs a peripheral enhancement was observed.

On the US assessment, the presence of one GLN was found in 26 cats. In 2 cats, two GLNs were seen. The GLN was not identified in 2 cats. All the identified GLNs were located in the omentum of the lesser curvature of the stomach. The probe was placed parallel to the spine immediately caudal to the xiphoid process, and then moved slightly to the left until an image with the liver to the left of the screen and the stomach to the right was obtained. The GLNs were identified in the fat tissue between the liver and the stomach. The nodes presented an ovoid shape and were heterogeneous, hypoechoic, or isoechoic in comparison to the surrounding fat tissue and with a thin hyperechoic halo. A hyperechoic central line was visible in the 33.33% of the cases (Figure 1).

Hepatic lymph nodes (HLNs): the presence of one HLN was found in 5 cadavers. They were located at the porta hepatis, ventral and slightly to the left of the portal vein.

On the CT images, the HLNs were seen in 22 cats. On precontrast images, the HLNs were isoattenuating (50.00%) or slight hypoattenuating (40.91%); after contrast administration they mainly showed a homogeneous contrast enhancement (86.36%), and less frequently a slightly heterogeneous (9.09%) or peripheral enhancement (4.55%). On the CT transverse images, these LNs were frequently identified dorsally and slightly to the right of the portal vein.

The frequency of identification of the HLNs on US images was similar to CT. The HLNs were visualized in 21 cats. These LNs showed more frequently a rounded shape (61.90%) and less frequently were elongated (33.33%) or miscellaneous (4.76%). Additionally, these LNs were most frequently isoechoic (42.86%) or hypoechoic (47.62%), and less frequently heterogeneous (9.52%). The HLNs were identified placing the probe parallel to the spine and caudally to the xiphoid process, and moving the probe

114 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

slightly to the right of the patient. The HLNs were seen adjacent to the portal vein while fanning slightly the probe right to left. A hyperechoic central line was visible in the 6.60% of the lymph nodes (Figure 2).

Splenic lymph node (SpLN): A single LN was identified in 2 cadavers. The SpLN was round in shape, and found embedded in the fat tissue of the splenic hilus.

On CT images, also a single SpLN was identified in 23 cats. The attenuation of the identified LNs showed the same frequency of isoattenuating (39.13%) and slightly hypoattenuating (39.13%) appearance. The remainder 21.74% of the LNs was classified as heterogeneous due to the presence of a hypoattenuating central area with soft tissue attenuation in the periphery. After contrast administration, 43.48% of the LNs presented homogenous contrast enhancement, 52.17% presented a peripheral contrast enhancement with a more hypoattenuating central area, and 4.35% were heterogeneous (Figure 3).

Ultrasonographically, this LN was identified also in 23 cats. The appearance of the SpLN was commonly hypoechoic (47.83%). However, a center isoechoic to the mesenteric fat within a hypoechoic periphery was identified in 34.78% of the SpLNs and these were classified as heterogeneous. An isoechoic and a hyperechoic appearance were seen in 8.70% of the LNs, respectively. A round shape was seen in 77.27% of SpLNs and, the remainder 22.73% was elongated. A hyperechoic central line was identified in 26.70% of the lymph nodes. The splenic vein was the main landmark to localize this LN, when present; the SpLN was adjacent to the vein in the fat tissue of the splenic hilus, close to the head of the spleen.

Pancreaticoduodenal lymph node (PdLN): was identified in all the cadavers, and was more often seen adjacent to the cranial pancreaticoduodenal vessels.

115 STUDIES

On the CT images, the PdLN was identified in 28 cats. This LN was most commonly slightly hypoattenuating (44.83%) or isoattenuating (27.59%) and less frequently heterogeneous (17.24%) or hypoattenuating (10.34%). After contrast administration, the contrast enhancement of the PdLN was variable, being homogeneous (31.03%), slightly heterogeneous (20.69%), heterogeneous (10.34%), or with a peripheral enhancement (37.93%).

On the US images, the PdLN was identified in 29 cats. A similar appearance in echogenicity to the described for the SpLN was observed for this LN. The PdLN was homogenously hypoechoic with a thin hyperechoic rim at the periphery in 37.93% of the cases. In 37.93% of the cases it presented a central area isoechoic to the mesenteric fat, surrounded by a hypoechoic ring and a thin hyperechoic rim at the periphery. This LN was found always on the right side, slightly caudal and ventral to the pylorus (figure 4).

Cranial mesenteric lymph center (lymphocentrum mesentericum craniale)

Jejunal lymph nodes (JLNs): in the anatomic study, 1 to 4 JLNs were identified. Three cadavers presented 4 JLNs. The presence of 3, 2, and 1 JLN was found in one cadaver each. These LNs were located along the jejunal vessels just proximal to the origin of the ileocolic artery. At least two large JLNs with a miscellaneous shape were identified in each animal. Smaller JLNs, rounded or elongated (Figure 5) were also present in some animals.

In the CT images, 1 to 3 JLNs were frequently identified. Three JLN were found in 15 cats. Two JLNs were present in 13 cats and only 1 JLN was seen in 2 cats. Their appearance was isoattenuating or slight hypoattenuating. More than one half of the JLNs presented a homogeneous contrast enhancement, the remainder presented slightly heterogeneous or heterogeneous contrast enhancement.

On the US assessment, 4 JLNs were present in 1 cat. The visualization of 3, 2, and 1 LNs was possible in 16, 11, and 1 cats, respectively. These LNs were mainly hypoechoic or isoechoic when compared with the surrounding fat

116 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

tissue. A hyperechoic central line was visible in only 5.20% of the LNs. The majority of the JLNs were classified as miscellaneous in shape. An elongated shape was less frequently observed.

Ileocecal lymph nodes (ICLNs): the presence of two ICLNs was observed in 4 cadavers, and only 1 ICLN was seen in 2 cadavers. They were located in the ileocecal ligament along the ileocecal vessels, one to each side of the cecum. These LNs were commonly rounded.

On the CT images, two ICLNs were identified in 26 cats, and only one in 2 cats. These LNs were located slightly caudal to the ileocolic junction always on the right side of the patient. On transverse slices, these LNs were identified in a dorsal and ventral position related to the ileocolic junction. The ICLNs were commonly isoattenuating or slight hypoattenuating, and less frequently heterogeneous. After contrast administration, the ICLNs frequently presented a homogeneous contrast enhancement, and with less frequency, these LNs showed slightly heterogeneous or heterogeneous contrast enhancement.

In the US assessment, two ICLNs were identified in 24 cats, and only one in 6 cats. These nodes were more frequently hypoechoic to isoechoic and less frequently heterogeneous in echogenicity. A hyperechoic central line was visible in 5.50% of the LNs. The ICLNs presented almost a similar distribution of shape, being rounded, elongated, or miscellaneous. The main landmark to find the ICLN was the ileocolic junction. An image of this with a sagittal plane of the ileum was obtained, and then a slight movement to the right of the patient allowed the visualization of these LNs (Figure 6).

Colic lymph nodes (CoLNs): 5 and 3 CoLNs were found in each of 2 cadavers. Additionally, the identification of 4 and 2 CoLNs was possible in one cadaver each. The CoLNs were located in the mesocolon, near the ascending and transverse colon. At least 2 LNs were bigger and more elongated than the rest, normally one near the ascending colon and one in a group of nodes located near the transverse colon.

117 STUDIES

On the CT images, the identification of 1, 2, 3, 4, and 5 CoLNs was possible in 12, 6, 4, 1, and 4 cats, respectively, for a total of 60 LNs in 27 cats. These LNs were commonly slight hypoattenuating or isottenuating. A homogeneous contrast enhancement was most frequently seen after contrast administration.

On US, only 1 CoLN could be successfully identified in 21 cats. This LN was close to the ileocolic junction. To differentiate it from the ICLNs the probe was displaced medially from the ileocolic junction instead of laterally as for the ICLNs. Frequently, the CoLNs were hypoechoic to the surrounding mesenteric fat, and less frequently isoechoic or heterogeneous. A hyperechoic central line was visible in 6.70% of these LNs. Almost half of these LNs were rounded, the rest were elongated or miscellaneous (figure 7).

Caudal mesenteric lymph center (lymphocentrum mesentericum caudale)

One to 4 caudal mesenteric lymph nodes (CMLNs) were identified in the cadavers. Two CMLNs were present in 1 cadaver. One, 3, and 4 CMLNs were found in 1 cadaver each. These LNs were located along the caudal mesenteric vessels and had an ovoid or rounded shape.

On CT images, one LN was seen in 14 cats; two, 3, and 4 CMLNs were found in 9, 3, and 2 cats, respectively. These LNs were commonly slightly hypoattenuating or isoattenuating. A homogeneous contrast enhancement was frequently seen after contrast administration. These LNs were seen on transverse images located slightly ventral to the descending colon and dorsolateral to the left aspect of the urinary bladder (figure 8).

On the US assessment, only one CMLN was identified in 10 cats. The CMLN was located between the bladder and the descending colon. This LN was commonly hypoechoic, with a thin hyperechoic rim and a fusiform elongated shape. A hyperechoic central line was visible in 3.30% of these LNs.

118 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Lumbar lymph center (lymphocentrum lumbale)

Lumbar aortic lymph nodes (LALNs): the presence of 1, 2, and 3 small rounded LNs were seen in one cadaver each. These LNs were located along and between the aorta and caudal vena cava.

On CT images, one LALN was found in 3 cats, and 2 LALN were found in 2 cats. The presence of fat tissue around the aorta and caudal vena cava in these 5 cats gave enough contrast to differentiate these LNs from the surrounding tissue. However, the localization of the LALN on CT images was dorsal to the aorta rather then between this and the caudal vena cava. In the other 25 cats that was not the case. The LALNs were very small, rounded and slightly hypoattenuating. A homogeneous contrast enhancement was present (Figure 9).

The LALNs were not identified in the US assessment.

Renal lymph nodes (RLNs): the left and right RLNs could be identified only in 2 cadavers. These small and rounded LNs were seen in close contact with the renal vessels.

On CT images, both RLNs were identified in 5 cats, and only the right RLN was seen in 2 cats. These LNs presented a thin soft tissue attenuating periphery surrounding a hypoattenuating center. After contrast administration, a peripheral contrast enhancement was observed. These LNs were located dorsal and slightly cranially to each renal vessel.

On the US images, only the right RLN was seen in 2 cats. These LNs were hypoechoic and ovoid in shape.

119 STUDIES

Iliosacral lymph center (lymphocentrum iliosacrale)

Medial iliac lymph nodes (MILNs): One right and one left MILNs were identified in 5 cadavers. These LNs were located slightly dorsal and lateral to the external iliac arteries and extended cranially until the deep circumflex iliac vessels. These LNs were commonly elongated but some were bilobed.

On the CT images, both right and left MILNs were identified in 28 cats. The location of these LNs was as described in the anatomic study, and more specifically extending from the cranial end-plate of the seventh lumbar vertebra (L7) until the cranial end-plate of the first sacral vertebra (S1). Some MILNs were very thin in their mid-portion presenting a bilobed shape. More than a half of these LNs were isoattenuating, the rest of them were slightly hypoattenuating compared with the surrounding muscles. The MILNs presented most frequently a homogeneous contrast enhancement.

Ultrasonographically, both right and left MILNs were identified in all the cats. In 50% of the cases, the right and left MILNs were isoechoic to the surrounding tissue, the other half were hypoechoic (R: 26.67%; L: 30.00%) or heterogeneous (R: 23.33%; L: 20.00%). A hyperechoic central line was identified in 18.30% of the LNs. The MILNs were fusiform in 83.33% of the cases and miscellaneous in 16.67% (Figure 10).

Internal iliac lymph nodes (InILN): a right and left InILNs were found only in 1 cadaver, along the internal iliac vessels, and commonly presented an elongated shape.

On the CT images, both right and left InILNs were visualized in 17 cats, additionally, one right and one left InILN were identified in each of 2 cats. These LNs were located at the ventrolateral aspect of S1 (ventral to the sacral wings) along the iliac internal vessels in contact with the body of the ilium. Most of these LNs were isoattenuating, and less frequently, slightly hypoattenuating or hypoattenuating. After contrast administration, these LNs showed homogeneous contrast enhancement (Figure 11).

120 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

On US assessment, these LNs were not identified.

Sacral lymph nodes (SaLNs): these LNs were not identified in the anatomic study. On CT images, the right and the left SaLNs were identified in 2 cats. Only the right was identified in 6 cats, and only the left was identified in 2 cats. These LNs were located in the midline at the ventral aspect of S1. Most frequently, the SaLNs were isoattenuating, and a slightly hypoattenuating appearence was less common. After contrast administration, the SaLNs presented homogeneous contrast enhancement.

On US assessment, in 1/30 (3.33%) cats, a right SaLN was visualized only in sagittal plane, making possible to obtained length and height measurement. A transverse image was not possible to obtain, therefore, the width of this LN is not reported. The LN was hypoechoic, with fusiform shape and was located caudally to the aorta trifurcation.

Inguinofemoral lymph center (lymphocentrum inguinofemorale)

Superficial inguinal lymph nodes (SILNs): in the anatomic study, the right and the left SILNs were identified in 4 cadavers. The location of these LNs was cranial to the inguinal canal in contact with the external pudendal vessels, always embedded in fat tissue. Commonly, the SILNs presented fusiform shape.

On CT images, the right and left SILNs were identified in 24 cats. Other frequencies of detection are as follows: one right SILN, one left SILN, and one right and two left SILN were seen in one cat each. These LNs were identified slightly caudal and dorsal to the junction between the external pudendal and caudal epigastric vessels, ventral to the pubic bone. Commonly they were slightly hypoattenuating and, less frequently, isoattenuating. After contrast administration, most of them showed homogeneous enhancement, and they

121 STUDIES were less frequently slightly heterogeneous or showed peripheral enhancement.

On the US assessment, the right and left SILNs were identified in 28 cats, and only the left SILN in 1 cat. Approximately half of these LNs were hypoechoic, the rest were heterogeneous (with more echogenic areas in the center of the LNs), isoechoic, and, less frequently hyperechoic. A hyperechoic central line was visible in 21.80% of these LNs (Figure 12).

Caudal epigastric lymph nodes (CELNs): this group of LNs was only identified in the assessment of the CT images. In 26 cats, one right and occasionally 3 (in 5/26) and one left and occasionally 2 (in 2/26) LNs were identified. Commonly these LNs were slightly hypoattenuating or isoattenuating in comparison with the muscle attenuation. After contrast administration, a homogeneous contrast enhancement was frequently seen; a peripheral or slightly heterogeneous contrast enhancement was less common. The location of these LNs was along the caudal epigastric vessels in the subcutaneous adipose tissue of the ventrolateral abdominal wall (at the level of L6-7) (Figure 13).

Ischiatic lymph center (lymphocentrum ischiadicum)

Ischiatic lymph nodes (IsLNs): the right and left IsLNs were found only in 1 cadaver. These rounded nodes were located at the base of the tail, partially covered by the gluteofemoralis muscle and embedded in adipose tissue.

On the CT images, both IsLNs were found in 1 cat. Additionally, five cats presented only the right IsLN, and 2 cats presented only the left IsLN. The location was as described in the anatomic study. These LNs were commonly rounded, isoattenuating, and less frequently slightly hypoattenuating and presented homogeneous contrast enhancement (Figure 14).

The IsLNs were not assessed with US.

122 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Popliteal lymph center (lymphocentrum popliteum)

The right and left (PoLNs) were identified in all the cadavers. These nodes were commonly rounded, and were embedded in the adipose tissue located in the caudal aspect of the stifle joint and in contact with the medial aspect of the lateral saphenous vein.

On CT images, both PoLNs were identified in 29 cats. These LNs commonly presented a hypoattenuating center surrounded by soft tissue attenuation. Additionally, approximately half of them were slightly hypoattenuating compared with muscle attenuation. After contrast administration, the PoLNs showed frequently a homogeneous contrast enhancement, and less commonly, peripheral enhancement.

On the US assessment, the right and left PoLNs were identified in 29 cats, and only the left was seen in 1 cat. Commonly rounded, these nodes were the only ones in the study with varied echogenicity among patients showing similar distribution of isoechoic, hypoechoic, hyperechoic, and heterogeneous echogenicity (Figure 15).

Statistical analysis

The mean and SD of the length, width, and height measurements obtained in the anatomic and imaging studies of each LN are summarized in Table 1.

Measurements data from CT, US, and anatomy showed differences when compared with the Wilcoxon signed rank test (CT and US) and with the Mann- Whitney U test (CT, US and Anatomy). Those differences are also shown in Table 1.

In this study, the MPR length on CT images showed a better understanding of the dimensions of the abdominal LNs, as previously mentioned in the author’s first report (Tobón Restrepo, et al. 2016), than the calculated length. Therefore, the MPR length was also chosen to compare between techniques

123 STUDIES even though only a few significant differences were observed in the comparison of both CT measurements. The comparison of the MPR length of the LNs showed few differences with US and anatomy. The length of the HLN obtained with US was significantly lower when compared with CT and anatomy. Additionally, the length of the JLNs was almost 10mm higher in CT than in US, and this difference was statistically significant. The length of both ICLNs in anatomy was lower compared with CT and US showing statistical significance. Although the length of the MILNs in CT was higher than in anatomy and US, only the length of the right MILNs showed statistically significant differences between CT and anatomy.

The width of the lymph nodes in the anatomic study was often lower than the measures obtained by CT and US. Statistical significant differences were observed for the PdLN, the second and third JLNs, the ICLNs, the first CoLN, and the left ISLN. The width of the GLN in US was higher compared with CT and anatomy and the differences were statistically significant. The width of the right MILN in anatomy was slightly higher than in CT, being statistically significant.

The height was the measurement that showed more statistically significant differences among techniques. The height of the GLN, the SpLN, the PdLN, the JLNs, the left SILN, and the ICLNs was higher when obtained with CT and US than in cadavers, and showed statistically significant differences among techniques; besides, the height in CT was also statistically significantly higher than in US. Additionally, the height of the HLN, the CoLN (except the most caudal), the first CMLN, and the right MILN was higher when obtained with CT than with anatomy and the differences were statistically significant.

The mean and SD of the Hounsfield units and the description of the attenuation for each group of lymph nodes is reported in Table 2. Most of the lymph nodes were isoattenuating or slightly hypoattenuating to surrounding musculature. A negative mean HU was obtained in the left RLN and LALN. However, these LNs were very small and most of them had a hypoattenuatting center and were classified as heterogeneous.

124 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

After contrast administration, most of the LNs showed homogeneous contrast enhancement. However, the RLNs (right: 57.14%; left: 75.00%) were the group that frequently presented a heterogeneous attenuation, followed by the SpLN (21.74%), and PoLNs (17.24%). A peripheral contrast enhancement was present in 8 lymph nodes, being more frequent in the SpLN.

A description of the ultrasonographic features of the LNs in this study is summarized in Table 3. The majority of the lymph nodes were hypoechoic followed by isoechoic. A small percentage showed a center isoechoic to the mesenteric fat with a more hypoechoic periphery and a hyperechoic rim; therefore they were classified as heterogeneous. Mainly the GLNs, the SpLN, the MILNs, the SILNs, and the PoLNs presented a higher percentage of heterogeneity than the other lymph nodes.

An elongated shape was commonly found in the LNs of this study, however, rounded (HLN, SpLN, PdLN and PoLNs) and miscellaneous (JLNs, MILNs, and ICLNs) shapes were also present.

Discussion

The identification of almost all the lymph centers of the abdominal and hindlimb regions was possible in plain dissection. However, an accurate assessment of some lymph nodes (e.g. LALNs, SaLNs, and CELNs) was not achieved because their similar appearance with the surrounding fat tissue. A dyeing procedure was not performed previous to the dissection because the animals were presented to post mortem examination within 24 hours of death. In our study the mean length of the abdominal LNs in the anatomic study was smaller than the reports in the previous literature. Sugimura, Kudo, & Takahata, (1958) reported a mean length range for the abdominal LNs of 1 to 8cm. In that study, Sugimura included cats from 7 days to 10 years old, resulting in a 33.3% of cats under 1-year-old. Previous studies reported statistical significant differences in the size of LNs when comparing young vs adult animals (Burns, Scrivani, Thompson, & Erb, 2008; Krol & O ’brien, 2012).

125 STUDIES

The abdominal and hindlimb lymph nodes in dogs have been well documented in the anatomy books, and assessed using CT and US (Beukers et al., 2013; Bezuidenhout, 2013; NAV, 2012; Nyman, Kristensen, Skovgaard, & McEvoy, 2005; Rossi, Patsikas, & Wisner, 2011). In cats, the anatomic locations, landmarks, drainage areas and number of LN per lymph center have been well documented (NAV, 2012; Saar & Getty, 1982; Tompkins, 1993), however, in those anatomic references, not all the measurements of lymph nodes were reported. In our study we report not only the length of the abdominal and hindlimb LNs but their width and height as well.

To the authors’ knowledge, this is the first report of the normal computed tomographic characteristics of the abdominal and hindlimb lymph nodes in the cat.

In one study from Schreurs et al. (2008), a description of the ultrasonographic characteristics of the LNs in the abdominal cavity was reported. In that study, Schreurs et al. reported the visualization of the MILNs, JLNs, HLNs, PdLN, SpLN, and LALNs with a frequency between 60 and 100% in the cat. In contrast, we found the ICLNs, CoLNs, and GLN also with a high frequency. Additionally, in our study the identification of the CMLNs was more frequent than in the Schreurs study. Differently from the Schreurs study, the LALNs, InILNs, left SaLN and left RLN were not found in our study. We hypothesize that the absence of identification of these LNs in our study, was due to the lack of contrast between these small structures and the tissues between the aorta and caudal vena cava. The use of real-time compound imaging US in Schreurs et al. study could be the reason for the differences with our study. This US modality produces a superior border definition and soft tissue contrast compared with B-Mode, and might have increased their ability to depict these LNs.

The popliteal LN was assessed in one study during ultrasound-guided intranodal injection of contrast medium, as part of a CT lymphography of the thoracic duct (Lee et al., 2012). The US characteristics were not provided but their measurements were similar to those in the present study.

126 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Few statistical differences were found regarding the length of the LNs among techniques. The HLN and the JLNs showed lower lengths in US that in the other techniques. We hypothesized that the presence of gas in the stomach and intestinal loops, the location (HLN), and a miscellaneous shape (JLN) could have influenced the achievement of a correct acoustic window to obtain the whole length of these LNs. Similar limitation have been reported in dogs (Agthe, Caine, Posch, & Herrtage, 2009).

The mean length of the LNs on US in our study showed some differences with those reported by Schreur et al. (2008). The HLN and CMLN in our study were longer. Additionally, the PdLN and right SaLN in our study were shorter. The cause for these differences remains unclear, it is possible that the sample size, fasting period of the cats, the scanning planes, and the interobserver variability were contributing factors.

The width and height of the LNs obtained in the anatomic study were shorter than in the imaging techniques. Some of them showed statistical differences, mainly for the PdLN, JLN, ICLN, and CoLN. We hypothesized that the lack of blood perfusion and the loss of fluids after dead could contribute to these differences.

The height was the most variable measurement between CT and US. A possible explanation is that the relative orientation of LNs is likely to be influenced by patient position, scanning plane used for the assessment, and filled intestinal loops and peristalsis that could induce displacement in dorso- ventral or lateral direction of the LNs. All the cats in this study were positioned in dorsal recumbency for the acquisition of the CT images. However, avoiding an oblique orientation of the abdominal lymph nodes in CT transverse images was challenging. Similar limitations have been described in dogs (Beukers et al., 2013).

The attenuation and Hounsfield Units of the LNs in this study before contrast administration was consistent with the descriptions and values reported for dogs (Beukers et al., 2013) and for the medial retropharyngeal LN in healthy cats (Nemanic & Nelson, 2012). In our study, some of the LNs in the abdominal cavity (SpLN, RLN) and in the hindlimb (PoLN) showed a nodal

127 STUDIES periphery isoattenuating or slightly hypoattenuating to the musculature, meanwhile the center of the node was hypoattenuating. This appearance has been reported for abdominal LNs in dogs (Beukers et al., 2013; Kneissl & Probst, 2007; Rossi et al., 2011), and for the sternal and axillary LNs in healthy cats (Tobón Restrepo et al. 2016), and it is produced by the presence of a fatty hilus. After contrast administration, a homogeneous contrast enhancement was more frequently visualized. In the LNs with a fatty hilus, a peripheral enhancement was observed. This phenomenon is due to the presence of vascularized nodal tissue in the periphery meanwhile the hilus had a less vascularized fat tissue.

The abdominal and hindlimb LNs were more frequently hypoechoic or isoechoic to the surrounding tissue. Similar descriptions have been reported for dogs (Agthe et al., 2009). The presence of a hyperechoic center, especially in the PoLNs, was most likely due to fat in the hilus (feature also visible in CT). This hyperechoic center was different than the described hyperechoic central line. This line was thin and well-defined, corresponding to the description of the hilus (D’Anjou, 2008; Nemanic & Nelson, 2012; Nyman & O’Brien, 2007; Tobón Restrepo, 2015).

The shape of the lymph nodes in this study is similar to that of the previous reports for dogs (Agthe et al., 2009; D’Anjou, 2008; Krol & O ’brien, 2012; Llabrés-Díaz, 2004; Nyman & O’Brien, 2007). The LNs that presented a rounded shape were small in size and had regular margins; the shape could be compared with the image in CT, being the same in all the cases. In previous reports, rounded lymph nodes in combination with increased size, loss of the hilus, and echotexture changes were suggestive of malignancy (De Swarte et al., 2011; Dennis, Kirberger, Barr, & Wrigley, 2010; Nyman et al., 2005). Such changes were not present in the rounded LNs in our study.

The present study included several limitations. The search of feline cadavers with cause of death other than neoplastic or inflammatory diseases was challenging resulting in a small sample size in the anatomic study. The use of ink solutions or other staining procedures in the cadavers were not performed, which could have assisted in the differentiation of the lymph nodes from the

128 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

surrounding fat tissue during dissection, especially in the sublumbar region and for the superficial lymph nodes. The animals used in the imaging study were assessed with US immediately after the CT scan. Therefore, the analysis of the CT images was not performed at the same time of the US examination, making an exact correlation in the number of LNs identified with both techniques challenging. In some of the patients, a period of apnea during the whole body CT scan was difficult to achieve, and movement artifact was present, especially in the cranial abdomen. This artifact reduced the identification of the LNs, mainly from the celiac lymph center. All the cats included in this study were fasted in order to avoid complications during anesthesia, however, some of them presented gas and feces in the colon that may have reduced the visualization of the LNs of the lumbar and iliosacral lymph centers.

In conclusion, the identification of the lymph nodes of the abdomen and hindlimb is possible with both imaging techniques. The length of the LNs is more challenging to assess with US than with CT when a multiplanar reconstruction is used. The LNs are more frequently isoattenuating to surrounding musculature in CT. However, some LNs can show a hypoattenuating center corresponding to a fatty hilus. Frequently, the abdominal and hindlimb LNs are hypoechoic or isoechoic. An elongated shape with regular margins was most frequently visible except for the JLNs, MILNs, and ICLNs that showed a miscellaneous shape and the HLN, SpLN, PdLN, and PoLNs that were rounded. To the authors’ knowledge, this is the first report of the CT features of the abdominal and hindlimb LNs. Also, this is the first study comparing the size of this group of LNs with cadavers, US, and CT examinations.

References

Agthe, P., Caine, A. R., Posch, B., & Herrtage, M. E. (2009). Ultrasonographic Appearance of Jejunal Lymph Nodes in Dogs Without Clinical Signs of Gastrointestinal Disease. Veterinary Radiology & Ultrasound, 50(2), 195–200.

129 STUDIES

Beukers, M., Vilaplana Grosso, F., & Voorhout, G. (2013). Computed Tomographic Characteristics of Presumed Normal Canine Abdominal Lymph Nodes. Veterinary Radiology & Ultrasound. 54(6), 610–617.

Bezuidenhout, A. J. (2013). The lymphatic system. In H. E. Evans & A. De Lahunta (Eds.), Miller’s anatomy of the dog (Forth edit., pp. 535 – 562). St. Louis, Mo : Elsevier.

Burns, G. O., Scrivani, P. V, Thompson, M. S., & Erb, H. N. (2008). Relation Between Age, Body Weight, and Medial Retropharyngeal Lymph Node Size in Apparently Healthy Dogs. Veterinary Radiology & Ultrasound Ultrasound, 49(3), 277–281.

D’Anjou, M.-A. (2008). Abdominal cavity, Lymph nodes, and Great Vessels. In D. Penninck & M.-A. D’Anjou (Eds.), Atlas of small animal ultrasonography (First edit., pp. 445 – 463). Ames : Iowa: Blackwell Publishing Ltd.

De Swarte, M., Alexander, K., Rannou, B., D’Anjou, M.-A. A., Blond, L., & Beauchamp, G. (2011). Comparison of sonographic features of benign and neoplastic deep lymph nodes in dogs. Veterinary Radiology and Ultrasound, 52(4), 451–456.

Dennis, R., Kirberger, R. M., Barr, F., & Wrigley, R. H. (2010). Handbook of small animal radiology and ultrasound : techniques and differential diagnoses (Second edit.). Edinburgh [etc.] : Elsevier.

Kneissl, S., & Probst, a. (2007). Comparison of computed tomographic images of normal cranial and upper cervical lymph nodes with corresponding E12 plastinated-embedded sections in the dog. Veterinary Journal, 174(2), 435–438.

Krol, L., & O ’brien, R. (2012). Ultrasonographic assessment of abdominal lymph nodes in puppies. Vet Radiol Ultrasound, 53(4), 455 – 458.

Lee, N., Won, S., Choi, M. M., Kim, J., Yi, K., Chang, D., Yoon, J. (2012). CT thoracic duct lymphography in cats by popliteal lymph node iohexol injection. Veterinary Radiology & Ultrasound, 53(2), 174–180.

Llabrés-Díaz, F. J. (2004). Ultrasonography of the medial iliac lymph nodes in the dog. Veterinary Radiology & Ultrasound, 45(2), 156–165.

Mattoon, J. S., Berry, C. R., & Nyland, T. G. (2015). Abdominal Ultrasound Scanning Techniques. In J. S. Mattoon & T. G. Nyland (Eds.), Small Animal Diagnostic Ultrasound (Third., pp. 94 – 127). St. Louis, Missouri : Elsevier.

NAV. (2012). Nomina anatomica veterinaria. ( Veterinary gross anatomical nomenclature International committee, Ed.) (Fifth Edit.). Oslo: ICVGAN.

Nemanic, S., & Nelson, N. C. (2012). Ultrasonography and noncontrast computed tomography of medial retropharyngeal lymph nodes in healthy cats. American Journal of Veterinary Research, 73(9), 1377–1385.

Nyman, H. T., Kristensen, A. T., Skovgaard, I. M., & McEvoy, F. J. (2005). Characterization of normal and abnormal canine superficial lymph nodes using gray-scale B-mode, color flow mapping, power, and spectral doppler ultrasonography: A multivariate study. Veterinary Radiology and Ultrasound, 46(5), 404–410.

Nyman, H. T., & O’Brien, R. T. (2007). The Sonographic Evaluation of Lymph Nodes. Clin Tech Small Anim Pract, 22(3), 128–137.

Rossi, F., Patsikas, M. N., & Wisner, E. R. (2011). Abdominal lymph nodes and lymphatic collecting system. In T. Schwarz & J. H. Saunders (Eds.), veterinary computed tomography (pp. 371 – 379). Ames : Iowa: Wiley-Blackwell.

Saar, L. I., & Getty, R. (1982). Sistema linfático de los carnívoros. In R. Getty (Ed.), S.Sisson - J.D.Grossman. Anatomía de los animales domesticos (Fifth edit., pp. 1811–1831).

130 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Barcelona, Spain : Masson

Schreurs, E., Vermote, K., Barberet, V., DAMINET, S., Rudorf, H., & Saunders, J. H. (2008). Ultrasonographic anatomy of abdominal lymph nodes in the normal cat. Veterinary Radiology & Ultrasound, 49(1), 68–72.

Sugimura, M., Kudo, N., & Takahata, K. (1958). Studies on the lymphonodi of cats: III. Macroscopical observations on the lymphonodi in the abdominal and pelvic cavities. Japanese Journal of Veterinary Research, 6(2), 69 – 88.

Tobón Restrepo, M. (2015). Ultrasound of the abdominal cavity, lymph nodes and large vessels. In R. Novellas Torroja, E. Dominguez Miño, Y. Espada Gerlach, Y. Martínez Pereira, & M. Tobón Restrepo (Eds.), Diagnostic ultrasound in cats (1st ed., pp. 211 – 229). Zaragoza: Spain: Servet editorial - Grupo Asís Biomedia S.L.

Tobón Restrepo, M., Espada, Y., Aguilar, A., Moll, X., & Novellas, R. (2016). Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb. Pending of publication.

Tompkins, M. B. (1993). Lymphoid system. In Atlas of feline anatomy for veterinarians (pp. 113 – 126). Philadelphia [etc.] : W.B. Saunders Company.

Widmer, W. R., Mattoon, J. S., & Nyland, T. G. (2015). Peritoneal Fluid, Lymph Nodes, Masses, Peritoneal Cavity, Great Vessels Thrombosis and Focused Examinations. In J. S. Mattoon & T. G. Nyland (Eds.), Small Animal Diagnostic Ultrasound (Third edit., pp. 501 – 516). St. Louis, Missouri : Elsevier Saunders.

Table 1. Mean (m) and SD for length (MPR, Calculated), width and height of the LNs from the abdomen and hindlimb in healthy cats. Comparison among techniques.

Length (mm) MPR CT-MPR CT-MPR US Lymph CT-MPR CT-Calc. US Anatomy vs vs vs vs node m (SD) m (SD) m (SD) m (SD) Calc. (A) US (A) Anatomy (B) Anatomy (B) G1 6.58 (3.18) 6.34 (2.62) 5.99 (2.05) 8.56 (6.45) G2 6.12 (3.91) 4.48 (0.74) 4.00 (0.28) - H 9.03 (2.77) 9.14 (3.26) 6.15 (2.32) 10.84 (2.17) * ** Sp 5.33 (2.01) 5.08 (1.63) 5.41 (2.03) 4.35 (2.62) Pd 6.38 (1.82) 7.20 (2.18) 5.76 (1.49) 5.60 (3.06) ** J1 36.97 (5.42) 34.50 (7.80) 25.05(6.19) 45.30 (10.83) * * J2 26.73 (4.75) 23.60 (7.20) 17.87 (7.74) 26.78 (12.45) * * J3 13.80 (2.50) 12.10 (3.40) 10.00 (2.34) 11.08 (1.50) * J4 - - - 18.67 (18.58) IC1 7.38 (2.21) 7.82 (2.52) 7.13 (2.65) 5.78 (2.37) * IC2 7.1 (1.96) 7.44 (2.76) 6.55 (1.92) 5.10 (0.88) * Co1 11.86 (3.71) 12.58 (4.00) 8.14 (2.39) 15.68 (10.59) * Co2 7.57 (2.51) 8.63 (2.49) - 8.75 (3.62) * Co3 5.98 (1.92) 6.88 (3.25) - 6.85 (1.45) Co4 4.06 (0.77) 4.76 (1.16) - 6.70 (3.07) Co5 5.20 (3.52) 5.80 (2.42) - 4.85 (2.89) CM1 10.34 (4.50) 11.37 (4.60) 10.00 (2.22) 7.34 (4.99) CM2 8.98 (4.44) 9.44 (4.27) - 7.52 (2.23) CM3 10.88 (6.26) 10.24 (5.01) - 6.80 (5.09) CM4 13.05 (0.49) 12.80 (0.42) - 6.70 (-) LA1 3.93 (1.55) 4.25 (1.87) - 3.57 (2.02) LA2 - - 5.40 (3.25) LA3 - - 4.50 (-) RR 9.30 (3.87) 9.20 (2.97) 8.10 (1.70) 11.40 (8.77) LR 8.80 (2.20) 9.25 (2.55) - 8.65 (4.60)

RMI 17.85 (4.85) 16.52 (4.78) 10.27 (3.56) 12.48 (6.58) ** LMI 17.52 (4.11) 17.04 (7.03) 11.62 (3.46) 13.28 (4.74) RInI 8.17 (2.65) 7.96 (2.14) - 12.00 (NC) LInI 8.66 (2.60) 8.08 (1.88) - 6.70 (NC) RSa 5.09 (2.48) 5.58 (2.24) 5.20 (NC) - LSa 4.65 (1.38) 5.65 (1.72) - - RSI 9.03 (3.17) 9.24 (3.03) 7.25 (1.75) 13.60 (5.62) LSI1 8.12 (3.23) 8.81 (4.14) 7.21 (2.28) 13.05 (5.33) LSI2 9.90 (NC) 6.90 - - RCE1 12.55 (6.98) 12.25 (5.56) - - RCE2 7.53 (5.16) 7.83 (3.87) - - RCE3 12.70 (NC) 13.80 (NC) - - LCE1 13.37 (5.38) 13.71 (5.35) - - LCE2 16.25 (1.63) 15.00 (0.85) - - RIs 4.58 (1.28) 4.92 (1.65) - 3.20 (-) LIs 4.63 (0.50) 5.63 (0.65) - 3.30 (-) RPo 8.02 (1.78) 8.24 (1.93) 7.73 (2.36) 7.37 (2.28) LPo 8.49 (2.30) 8.14 (2.35) 7.88 (2.35) 6.63 (1.82) * p-value < 0.05 (A) Wilcoxon Signed Rank Test ** p-value < 0.01 (B) Mann-Whitney U Test

Continuation Table 1. Mean (m) and SD for length (MPR, Calculated), width and height of the LNs from the abdomen and hindlimb in healthy cats. Comparison among techniques.

Width Height (mm) (mm) CT CT US CT CT US Lymph CT US Anatomy CT US Anatomy vs vs vs vs vs vs node m (SD) m (SD) m (SD) m (SD) m (SD) m (SD) US (A) Anatomy (B) Anatomy (B) US (A) Anatomy (B) Anatomy (B) G1 3.30 (1.06) 4.87 (1.16) 3.14 (1.05) ** * 3.04 (1.27) 2.50 (0.72) 1.1 (0.14) * ** ** G2 3.23 (1.11) 3.50 (2.40) - 3.43 (0.86) 3.10 (0.28) - H 3.89 (1.67) 5.33 (1.95) 4.20 (2.02) 4.48 (1.62) 2.92 (0.77) 2.44 (1.55) * Sp 4.12 (1.32) 3.80 (1.31) 2.65 (1.20) 3.55 (1.38) 2.53 (0.94) 1.5 (0.71) * * Pd 6.05 (1.79) 5.77 (2.13) 2.55 (1.05) ** ** 4.31 (1.36) 3.17 (0.94) 1.1 (0.64) ** ** ** J1 9.55 (4.44) 10.93 (5.23) 7.90 (3.50) 6.27 (2.58) 4.42 (1.55) 2.50 (1.87) * * J2 9.58 (4.91) 15.27 (6.04) 5.40 (1.31) * * * 5.78 (1.70) 4.23 (1.07) 1.55 (1.05) * * * J3 10.25 (3.06) 8.68 (4.69) 3.75 (0.71) * * 8.53 (4.11) 4.38 (1.75) 1.55 (0.24) * * * J4 - - 3.67 (1.80) - - - IC1 5.37 (1.54) 6.36 (2.84) 2.95 (0.91) * * 3.80 (1.20) 3.06 (1.01) 1.35 (0.39) ** ** ** IC2 7.57 (9.06) 5.96 (2.32) 3.45 (1.05) ** ** 6.41 (9.8) 2.94 (0.59) 1.47 (0.57) * ** ** Co1 5.94 (3.02) 7.44 (3.38) 3.05 (0.63) * * 5.35 (2.98) 3.68 (1.29) 1.23 (0.46) * Co2 5.09 (1.91) - 4.08 (1.57) 3.70 (1.53) - 1.35 (0.49) * Co3 4.58 (1.50) - 3.48 (1.12) 2.87 (0.92) - 1.30 (0.47) * Co4 5.40 (1.79) - 3.40 (0.89) 3.36 (1.47) - 1.20 (0.18) * Co5 4.20 (1.68) - 3.55 (1.77) 3.30 (1.00) - 1.45 (0.21) CM1 5.14 (2.21) 7.77 (2.69) 3.12 (0.73) 3.51 (1.44) 3.50 (0.98) 1.66 (0.64) * CM2 5.37 (2.36) - 3.12 (0.75) 3.53 (1.55) - 1.82 (1.20) CM3 6.38 (2.22) - 4.00 (1.70) 3.62 (1.82) - 1.90 (0.42) CM4 6.45 (2.05) - 4.00 (NC) 4.30 (1.27) - 1.90 (NC) LA1 2.17 (0.92) - 1.70 (0.72) 1.82 (1.13) - 0.87 (0.58) LA2 - - 1.95 (0.35) - - 1.30 (NC) LA3 - - 2.00 (NC) - - 0.90 (NC) RR 4.64 (1.61) 6.10 (2.69) 3.10 (0.42) 3.31 (1.07) 3.25 (0.64) 1.15 (0.21) LR 3.30 (1.29) - 2.30 (NC) 2.95 (1.38) - 1.30 (0.42)

RMI 3.46 (1.25) 4.64 (1.72) 3.82 (1.08) ** 2.42 (0.65) 3.45 (1.34) 1.40 (0.87) * LMI 3.31 (1.00) 4.12 (1.33) 2.88 (0.73) 2.60 (1.16) 3.28 (0.84) 1.56 (0.88) RInI 4.25 (1.91) - 4.70 (NC) 3.10 (1.50) - 2.50 (NC) LInI 4.07 (1.81) - 4.40 (NC) 3.73 (1.71) - - RSa 4.76 (1.83) - - 2.09 (0.67) 2.00 (NC) - LSa 4.72 (1.72) - - 2.17 (0.84) - - RSI 6.88 (2.53) 6.19 (3.11) 3.45 (1.69) 3.21 (1.14) 2.72 (0.90) 1.27 (0.72) LSI1 6.76 (2.95) 5.43 (1.82) 3.27 (1.49) * * 3.04 (1.28) 2.67 (0.62) 1.27 (0.70) ** ** LSI2 6.90 (NC) - - 4.20 (NC) - - RCE1 5.80 (1.96) - - 3.88 (1.25) - - RCE2 4.30 (0.82) - - 3.83 (1.95) - - RCE3 6.90 (NC) - - 5.90 (NC) - - LCE1 5.72 (1.95) - - 3.83 (1.35) - - LCE2 5.60 (2.4) - - 5.00 (0.85) - - RIs 4.05 (0.76) - 2.00 (NC) 2.42 (0.63) - 1.00 (NC) LIs 3.73 (0.91) - 6.00 (NC) 2.50 (0.1) - 1.00 (NC) RPo 6.29 (1.65) 5.99 (1.26) 4.45 (2.52) 5.11 (1.83) 4.99 (1.24) 1.58 (0.93) LPo 5.97 (1.25) 6.05 (1.32) 4.10 (1.80) 5.42 (1.45) 5.11 (1.36) 1.67 (1.04) G: gastric LN; H: hepatic LN; Sp: splenic LN; Pd: pancreaticoduodenal LN; J: jejunal LNs; IC: iliocecal LNs; Co: colic LNs; CM: caudal mesenteric LNs; LA: lumbar aortic LNs; RR: right renal LN; LR: left renal LN; RMI: right medial iliac LN; LMI: left medial iliac; RInI: right internal iliac LN; LInI: left internal iliac; RSa: right sacral LN; LSa: left sacral LN; RSI: right superficial inguinal LNs; LSI: left superficial inguinal LNs; RCE: right caudal epigastric LNs; LCE: left caudal epigastric LNs; RIs: right ischiatic LN; LIs: left ischiatic LN; RPo: right popliteal LN; LPo: left popliteal LN; NC: not calculated.

Table 2. Computed tomography characteristics of the lymph nodes of the abdomen and hindlimb of healthy cats.

Lymph HU Precontrast HU Postcontrast Attenuation Precontrast (%) Attenuation Postcontrast (%)

node Mean SD Min Max Mean SD Min Max Iso S Hypo Hypo Hyper Heter Hom S Heter Heter Peri

G1 20.66 26.18 -55.33 63.33 94.65 30.96 20.33 151.33 50.00 42.86 3.57 0.00 3.57 82.14 3.57 0.00 14.29 G2 33.05 22.91 -12.00 46.33 102.95 9.62 90.67 114.67 66.67 33.33 0.00 0.00 0.00 100.00 0.00 0.00 0.00 H 35.36 10.73 13.67 50.33 122.30 25.31 66.33 170.00 50.00 40.91 4.55 0.00 4.55 86.36 9.09 0.00 4.55 Sp 5.81 32.72 -64.00 50.33 79.83 38.47 -0.67 125.33 39.13 39.13 0.00 0.00 21.74 43.48 0.00 4.35 52.17 Pd 27.56 21.91 -29.33 55.00 100.15 35.99 -6.67 148.33 27.59 44.83 10.34 0.00 17.24 31.03 20.69 10.34 37.93 J1 38.76 9.24 18.33 52.00 143.26 24.31 94.33 170.67 60.00 33.33 0.00 3.33 3.33 50.00 40.00 10.00 0.00 J2 41.17 7.93 30.67 57.67 145.95 23.12 98.00 171.00 64.33 25.00 0.00 3.66 7.10 53.60 35.71 10.71 0.00 J3 37.67 7.82 19.67 48.33 122.45 13.52 101.67 146.33 57.10 42.90 0.00 0.00 0.00 64.30 35.70 0.00 0.00 IC1 38.15 10.37 18.67 55.33 108.03 21.33 72.33 148.00 48.15 40.74 0.00 3.70 7.41 81.48 11.11 7.41 0.00 IC2 40.84 11.19 11.67 56.00 106.17 21.61 64.33 139.33 51.85 44.44 0.00 0.00 3.70 74.07 11.11 11.11 3.70 Co1 38.30 13.75 8.00 71.67 118.40 21.04 64.00 152.33 51.90 48.10 0.00 0.00 0.00 85.20 14.80 0.00 0.00 Co2 36.31 10.24 20.00 54.33 123.67 24.85 55.33 152.67 40.00 53.30 0.00 0.00 6.70 86.70 0.00 13.30 0.00 Co3 31.37 15.39 7.33 54.00 126.52 21.11 94.33 166.00 44.40 55.60 0.00 0.00 0.00 100.00 0.00 0.00 0.00 Co4 30.33 18.52 -2.00 44.33 129.67 20.90 108.33 161.00 20.00 80.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 Co5 34.33 11.41 22.33 45.67 127.00 27.83 91.33 157.33 25.00 75.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 CM1 41.07 9.63 19.00 58.67 116.44 23.65 68.67 163.00 41.38 48.28 0.00 0.00 10.34 82.76 13.79 3.45 0.00 CM2 43.00 9.79 23.33 59.33 118.10 24.00 82.67 163.67 35.71 57.14 0.00 0.00 7.14 78.57 14.29 7.14 0.00 CM3 36.80 12.26 15.67 47.67 115.80 13.15 98.67 128.33 20.00 80.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 CM4 51.66 10.37 44.33 59.00 130.66 30.17 109.33 152.00 100.00 0.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 LA1 -4.25 15.95 -23.33 10.67 61.00 21.34 37.33 89.00 25.00 75.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 RR 12.67 36.33 -33.67 60.00 76.71 30.75 33.00 115.33 42.86 0.00 0.00 0.00 57.14 0.00 0.00 14.29 85.71 LR -4.50 18.91 -26.00 13.00 66.42 46.41 13.33 106.00 25.00 0.00 0.00 0.00 75.00 0.00 0.00 0.00 100.00 RMI 30.48 20.98 -18.33 56.67 109.80 33.54 46.33 166.33 57.14 42.86 0.00 0.00 0.00 92.86 3.57 3.57 0.00 LMI 37.62 17.11 -12.33 67.33 106.57 29.94 20.00 160.67 53.57 42.86 3.57 0.00 0.00 96.43 3.57 0.00 0.00 RInI 37.77 18.20 -3.33 58.00 120.25 24.10 80.33 165.00 68.42 26.32 5.26 0.00 0.00 73.68 21.05 0.00 5.26 LInI 38.25 20.71 -15.00 67.00 118.19 22.01 90.00 180.00 73.68 21.05 5.26 0.00 0.00 73.68 21.05 0.00 5.26 RSa 46.38 14.28 25.00 66.00 104.92 25.76 58.67 140.33 75.00 25.00 0.00 0.00 0.00 75.00 25.00 0.00 0.00 LSa 46.66 28.41 4.33 65.33 106.08 35.02 54.33 131.00 75.00 25.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 RSI 26.41 21.87 -37.00 60.00 99.48 35.05 11.00 169.67 35.71 64.29 0.00 0.00 0.00 92.86 3.57 0.00 3.57 LSI1 28.83 19.46 -34.67 61.00 95.57 32.81 -7.67 148.33 32.14 67.86 0.00 0.00 0.00 92.86 3.57 0.00 3.57

LSI2 54.00 NC 54.00 54.00 98.67 NC 98.67 98.67 0.00 100.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 RCE1 23.61 16.95 -22.67 50.00 106.91 38.95 35.00 206.67 34.62 57.69 0.00 0.00 7.69 76.92 11.54 0.00 11.54 RCE2 29.58 8.92 19.67 41.33 121.00 60.59 77.00 208.33 25.00 75.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 RCE3 24.67 NC 24.67 24.67 132.33 NC 132.33 132.33 0.00 100.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 LCE1 25.04 17.49 -32.00 49.00 112.68 37.10 26.33 178.67 30.77 61.54 0.00 0.00 7.69 80.77 7.69 0.00 11.54 LCE2 35.70 1.41 34.70 36.70 119.15 16.76 107.30 131.00 50.00 50.00 0.00 0.00 0.00 100.00 0.00 0.00 0.00 RIs 39.11 14.44 19.33 55.00 95.94 15.10 77.00 113.00 66.67 33.33 0.00 0.00 0.00 100.00 0.00 0.00 0.00 LIs 29.56 7.90 24.67 38.67 102.00 27.13 86.00 133.33 66.67 33.33 0.00 0.00 0.00 100.00 0.00 0.00 0.00 RPo 17.00 21.50 -32.67 45.00 91.12 36.12 -20.00 149.33 10.34 48.28 24.14 0.00 17.24 62.07 10.34 3.45 24.14 LPo 14.99 21.03 -34.33 47.67 92.30 32.39 14.33 171.67 6.90 51.72 24.14 0.00 17.24 55.17 10.34 3.45 31.03 G: gastric LN; H: hepatic LN; Sp: splenic LN; Pd: pancreaticoduodenal LN; J: jejunal LNs; IC: iliocecal LNs; Co: colic LNs; CM: caudal mesenteric LNs; LA: lumbar aortic LNs; RR: right renal LN; LR: left renal LN; RMI: right medial iliac LN; LMI: left medial iliac; RInI: right internal iliac LN; LInI: left internal iliac; RSa: right sacral LN; LSa: left sacral LN; RSI: right superficial inguinal LNs; LSI: left superficial inguinal LNs; RCE: right caudal epigastric LNs; LCE: left caudal epigastric LNs; RIs: right ischiatic LN; LIs: left ischiatic LN; RPo: right popliteal LN; LPo: left popliteal LN. Iso: isoattenuating; S Hypo: slightly hypoattenuating; Hypo: hypoattenuating; Hyper: hyperattenuating. Hom: homogeneous; S Het: slightly heterogeneous; Het: heterogeneous; Peri: peripheral enhancement. NC: not calculated.

Table 3. Ultrasonographic features of the lymph nodes of the abdomen and hindlimb in healthy cats.

Lymph Echogenicity (%) Shape (%)

node Isoechoic Hypoechoic Hyperechoic Heterogeneous Rounded Elongated Miscellaneous

G1 37.04 22.22 0.00 40.74 29.63 66.67 3.70 G2 0.00 50.00 0.00 50.00 50.00 50.00 0.00 H 42.86 47.62 0.00 9.52 61.90 33.33 4.76 Sp 8.70 47.83 8.70 34.78 77.27 22.73 0.00 Pd 24.14 37.93 0.00 37.93 51.72 48.28 0.00 J1 20.00 63.33 3.33 13.33 0.00 6.90 93.10 J2 28.60 57.10 3.60 10.70 0.00 14.29 85.71 J3 13.30 73.30 6.70 6.70 0.00 26.70 73.30 IC1 30.00 66.67 0.00 3.33 36.67 40.00 23.33 IC2 32.00 64.00 0.00 4.00 24.00 48.00 28.00 Co 19.05 71.43 0.00 9.52 42.86 28.57 28.57 CM 20.00 70.00 0.00 10.00 10.00 50.00 40.00 LA* ------RR 0.00 100.00 0.00 0.00 0.00 100.00 0.00 LR* ------RMI 50.00 26.67 0.00 23.33 0.00 83.33 16.67 LMI 50.00 30.00 0.00 20.00 0.00 83.33 16.67 RInI* ------LInI* ------RSa 0.00 100.00 0.00 0.00 0.00 100.00 0.00 LSa * ------RSI 28.57 42.86 3.57 25.00 10.71 85.71 3.57 LSI 24.14 55.17 3.45 17.24 17.24 82.76 0.00 RCE* ------LCE* ------RIs* ------LIs* ------RPo 33.33 16.67 20.00 30.00 93.33 3.33 3.33 LPo 37.93 13.79 24.14 24.14 96.55 3.45 0.00 * LNs that were not identified in US. G: gastric LN; H: hepatic LN; Sp: splenic LN; Pd: pancreaticoduodenal LN; J: jejunal LNs; IC: iliocecal LNs; Co: colic LNs; CM: caudal mesenteric LNs; LA: lumbar aortic LNs; RR: right renal LN; LR: left renal LN; RMI: right medial iliac LN; LMI: left medial iliac; RInI: right internal iliac LN; LInI: left internal iliac; RSa: right sacral LN; LSa: left sacral LN; RSI: right superficial inguinal LNs; LSI: left superficial inguinal LNs; RCE: right caudal epigastric LNs; LCE: left caudal epigastric LNs; RIs: right ischiatic LN; LIs: left ischiatic LN; RPo: right popliteal LN; LPo: left popliteal LN.

138 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Figure 1. Gastric lymph node. a. Image of the dissection showing its localization (arrow) in the lesser omentum (LO) of the stomach (S), near the pylorus (P). b. Ultrasonographic image showing an elongated gastric LN (between cursors), with an isoechoic center and a hypoechoic periphery, located between the stomach (S) and the liver (L). c – d. CT images indicating the localization of an isoattenuating gastric LN (arrow) in the lesser omentum in the precontrast image (c) and with a homogeneous contrast enhancement pattern in the postcontrast image (d). A second LN (arrow head) is partially visible dorsally, between the stomach (S) and the spleen (Sp). The liver (L) and pylorus (P) are indicated.

STUDIES 139

Figure 2. Hepatic lymph node. a. Image of the dissection showing the localization of the hepatic LN (arrow) in the porta hepatis, near to the portal vein (P). The liver (L) and gallbladder (GB) are indicated. b. Ultrasonographic image showing the hepatic LN (between cursors), in the porta hepatis. The liver (L) and a partially visible portal vein (P) are indicated. c – d. CT images indicating a slightly hypoattenuating hepatic LN (arrow) in the precontrast image (c) and with homogeneous contrast enhancement in the postcontrast image (d). It is located normally at the dorsomedial aspect of the portal vein (P), and ventromedial to the caudal vena cava (C). The liver (L) and the spleen (Sp) are indicated.

140 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Figure 3. Splenic lymph node. a. Image of the dissection showing its localization (arrow) in the splenic hilus, along the splenic vein (SV). The spleen (Sp) is partially visible. b. US image showing the splenic LN (between cursors), at the splenic hilus. The spleen (Sp) and a partially visible splenic vein (asterisk) are indicated. c – d. CT images indicating the localization of the splenic LN (short arrow) seen slightly hypoattenuating with a hypoattenuating center in the precontrast image (c) and showing slightly heterogeneous contrast enhancement (d). The splenic vein (long arrow), the pancreas (P), and the spleen (Sp) are indicated.

STUDIES 141

Figure 4. Pancreaticoduodenal lymph node. a. Image of the dissection showing the localization of the pancreaticoduodenal LN (arrow) in relation to the pancreas (P) and the duodenum (D). b. US image showing a heterogeneous, rounded pancreaticoduodenal LN (between cursors), between the duodenum (D), pancreas (P), and the liver (L). c – d. CT images indicating the localization of a rounded pancreaticoduodenal LN (arrow) seen isoattenuating with a hypoattenuating center in the precontrast image (c) and with peripheral homogeneous contrast enhancement in the postcontrast image (d). The liver (L), gall bladder (GB), duodenum (D), portal vein (asterisk), stomach (S), and spleen (Sp) are indicated.

142 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Figure 5. Jejunal lymph nodes. a. Image of the dissection showing the localization of the jejunal LN (arrow) in the mesentery along the jejunal vessels (JV). The jejunal loops (J) are indicated. b. US image showing a miscellaneous shaped, hypoechoic jejunal LN (between cursors). The jejunal vessels (asterisk) are partially seen. c – d. CT images in dorsal reconstruction indicating the localization of the jejunal LNs (arrows) seen isoattenuating in the precontrast image (c) and with homogeneous contrast enhancement in the postcontrast image (d). The jejunal loops (J), colon (Co), stomach (S), spleen (Sp), pancreas (P), and urinary bladder (UB) are indicated.

STUDIES 143

Figure 6. Ileocecal lymph nodes. a. Image of the dissection showing the localization of the ileocecal LNs (arrow) near the ileocolic junction, along the ileocolic vessels (asterisk). The ileum (I), cecum (Cc), and colon (Co) are indicated. b. US image showing 2 elongated and hypoechoic ileocecal LNs (between cursors) between the ascending colon (Co) and ileum (I). c – d. CT images indicating the localization of the ileocecal LN (arrows) seen isoattenuating in the precontrast image (c) and with homogeneous enhancement in the postcontrast image (d). The colon (Co), Ileum (I), jejunal loops (J), right (RK) and left (LK) kidneys, and spleen (Sp) are indicated.

144 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Figure 7. Colic lymph node. a. Image of the dissection showing the localization of the colic LN (arrow) near the ascending colon (Co) and ileum (I). b. Doppler US image showing the colic LN (among cursors), blood flow can be seen in the mesenteric vessels. c – d. CT images indicating the localization of the colic LN (arrow) seen isoattenuating in the precontrast image (c) and with homogeneous enhancement in the postcontrast image (d). The colon (Co), Ileum (I), jejunal loops (J), right (RK) and left (LK) kidneys, and spleen (Sp) are indicated.

STUDIES 145

Figure 8. Caudal mesenteric lymph node. a. Image of the dissection showing the localization of the caudal mesenteric LN (arrow) near the descending colon (Co) along the caudal mesenteric vessels (CM). b. US image showing the caudal mesenteric LN (between cursors), near the descending colon (Co). c – d. CT images indicating the localization of the caudal mesenteric LN (arrow) seen isoattenuating in the precontrast image (c) and with homogeneous enhancement in the postcontrast image (d). The descending colon (Co) and the urinary bladder (UB) are indicated.

146 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Figure 9. Lumbar aortic lymph nodes. a. Image of the dissection showing the localization of the lumbar aortic LN (arrow) between to the abdominal aorta (Ao) and the caudal vena cava (asterisk). The psoas (Ps) muscles are indicated. b - d. CT images indicating the localization of the lumbar aortic LN (arrow) in sagittal plane seen with heterogeneous contrast enhancement with a hypoattenuating central area postcontrast (b) and in transverse plane seen slightly hypoattenuating in the precontrast image (c) and with homogeneous enhancement in the postcontrast image (d). The psoas (Ps) muscles, aorta (Ao), caudal vena cava (asterisk), and right (RK) and left (LK) kidneys are indicated.

STUDIES 147

Figure 10. Medial iliac lymph nodes. a. Image of the dissection showing the localization of the medial iliac LNs (arrow) along the aortic trifurcation (asterisk) and the external iliac vessels (EI). The descending colon (Co) is indicated. b. US image showing an isoechoic medial iliac LN (between cursors) with fusiform shape located ventral to the psoas (Ps) muscles. c - d. CT images indicating the localization of the medial iliac LNs (arrows) in dorsal plane seen isoattenuating in the precontrast image (c) and with homogeneous enhancement in the postcontrast image (d), around the aortic trifurcation (asterisk) between the external iliac vessels (EI) and the psoas (Ps) muscles.

148 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Figure 11. Internal iliac lymph nodes. a - b. CT transverse images indicating the localization of isoattenuating internal iliac LNs (arrows) in the precontrast image (a) that show homogeneous enhancement in the postcontrast image (b). The LNs are located near the medial aspect of the ilium (I), ventral to the sacrum (S1) and along the internal iliac vessels (asterisk). The colon (Co) and the urinary bladder (UB) are indicated.

STUDIES 149

Figure 12. Superficial inguinal lymph nodes. a. Image of the dissection showing the localization of the superficial inguinal LNs (arrow) along the external pudendal vessels (asterisks). b. US image showing a heterogeneous superficial inguinal LN (between cursors) embedded in the inguinal adipose tissue (asterisks). c - d. CT transverse images indicating the localization of the superficial inguinal LNs (arrows) seen isoattenuating in the precontrast image (c) and with homogeneous enhancement in the postcontrast image (d). The pubis (asterisk) is indicated.

150 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Figure 13. Caudal epigastric lymph nodes. a – d. CT transverse (a &b) and dorsal (c & d) images indicating the localization of the caudal epigastric LNs (arrows) in precontrast (a & c) and postcontrast (b & d) images, along the caudal epigastric vessels (asterisk). In a & b, mammary tissue is indicated (arrow heads).

STUDIES 151

Figure 14. Ischiatic lymph node. a - b. Images of the dissection showing the localization of the ischiatic LN (arrow) deep to the gluteofemoralis muscle (GF). c - d. CT transverse images indicating the localization of the ischiatic LN (arrow) seen isoattenuating in the precontrast image (c) and with homogeneous enhancement in the postcontrast image (d), deep to the gluteofemoralis (asterisk) muscle. A coccygeal vertebra (Cc) is indicated.

152 Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb

Figure 15. Popliteal lymph nodes. a. Image of the dissection showing the localization of the popliteal LN in the caudo-proximal aspect of the stifle joint (arrow). b. US image showing a heterogeneous (hyperechoic central area compatible with the hilus) popliteal LN (between cursors) caudally to the gastrocnemius muscle (G). c - d. CT images indicating the localization of the popliteal LN (arrows) seen heterogeneous with isoattenuating periphery and hypoattenuating center in the precontrast image (c) and with homogeneous peripheral enhancement in the postcontrast image (d). The gastrocnemius (G) muscles are indicated.

4.3. Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography.

155 STUDIES

Abstract

The lymph nodes (LNs) of a group of diseased cats (case group) were prospectively assessed with computed tomography (CT) and ultrasound (US) and compared with the CT and US images of the LNs from a control group. The length, height, short / long axis ratio, Hounsfield units (HU), attenuation, echogenicity, shape, and margins of the LNs were compared to investigate differences between normal and abnormal LNs, and between CT and US diagnostic accuracy. Three regions of study were determined according to the localization of primary disease as follows: R1 (head and neck), R2 (thorax and forelimb), and R3 (abdomen and hindlimb). The LNs of the cats from the case group were divided into inflammatory or neoplastic category according to the results from the LNs’ or primary lesion’s cytology or biopsy. Thirty cats were recruited in the control group. For the analysis of each region, the lymph centers (LC) of the control group were matched with the affected lymph centers from the 29 cats included in the case group. In the R1, 62 (I= 14; N= 48) LNs on CT and 58 (I= 14; N= 44) LNs on US were included, representing at least one LN from the mandibular and retropharyngeal lymph centers. In the R2, 22 (I= 11; N= 11) LNs on CT and 18 (I= 8; N= 10) LNs on US were included, mainly from the axillary, cranial mediastinal, and ventral thoracic lymph centers. In the R3, 59 (I= 19; N= 40) LNs on CT and 69 (I= 19; N= 50) LNs on US were included, representing at least one LN from the celiac, cranial and caudal mesenteric lymph centers. The quantitative variables, length, height, and S/L ratio, obtained with both techniques, and the HU before contrast administration, were significantly different between the control and the case group for the R2 and R3. In the R1 the significant differences were found only for the length and height obtained with US, and the HU before and after contrast administration. The distribution of the qualitative variables (attenuation pre- and postcontrast, echogenicity, shape, and margins) among groups was significantly different in the case group compared with the control group. The heterogeneous and hypoattenuating appearance of the most neoplastic and some inflammatory LNs before and after contrast administration could suggest severe changes in the internal structure and vascularization but a certain degree of malignancy

156 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

was not possible to determine with CT. A better assessment and more precise measurements were obtained with CT than with US, especially in the R2 (thorax), due to the absence of the artifacts that were present in US. In conclusion, the results of this study allowed the identification of some features of the lymph nodes that are suggestive of abnormality. However, a differentiation between inflammatory and neoplastic processes is more challenging due to overlapping of the features in both techniques.

Introduction

With the new technologies, it is possible to acquire excellent quality images with ultrasonography (US) and with computed tomography (CT), not only in human medicine, but also in veterinary medicine (De Swarte et al., 2011; Nemanic, Hollars, Nelson, & Bobe, 2015; Nyman & O’Brien, 2007; Wunderbaldinger, 2006). This current situation has lead to increase research in order to understand the features of the normal lymph nodes and the changes observed in diseases (Henninger, 2003; Li et al., 2013; Nemanic et al., 2015; Nyman, Kristensen, Skovgaard, & McEvoy, 2005; Nyman & O’Brien, 2007; Salwei, O’Brien, & Matheson, 2005). One of the primary goals in the evaluation of the lymph nodes is to determine if the changes observed are correlated with histological changes, so this could lead to an early and less invasive diagnosis, staging, and prognosis of diseases (Nyman & O’Brien, 2007; Wunderbaldinger, 2006)

The lymph nodes react to endogenous and exogenous agents with a variety of specific morphological and functional responses. Histological changes observed in lymph nodes include necrosis, apoptosis, sinus ectasia, vascular lesions, amyloidosis, lymphadenitis, hyperplasia (e.g. immune, reactive, plasma cell), extramedullary hematopoiesis, lymphoma, and metastases (Elmore, 2006).

Morphologic changes of the lymph nodes (LNs) may be identified using US or CT images. Variations in size, shape, indistinct nodal margins, parenchymal heterogeneity, distinct distribution pattern of the vessels, variations in the

157 STUDIES resistive and pulsatility indices, and loss of perinodal fat and hilus have been commonly reported in pathologic lymph nodes (Gendler, Lewis, Reetz, & Schwarz, 2008; Karnik, Reichle, Fischetti, & Goggin, 2009; Nemanic et al., 2015; Nyman & O’Brien, 2007; Tobón Restrepo et al., 2015). However, correlations with the findings in US and/or CT and a specific histologic change or disease in cats are needed.

The aims of the study were; (i) to investigate potential differences in the LNs features evaluated with CT and US between healthy cats and cats with disease, and (ii) to assess the ability of each imaging technique (CT and US) to discriminate between neoplastic and inflammatory changes in the lymph nodes.

Material and methods

This study was approved by the ethical committee of the Universitat Autònoma de Barcelona (reference number CEAAH 2255 of September 2013). Written owner consent was obtained prior inclusion of the cats in this study.

For the purpose of this study, the animals were divided into two groups, a control group and a case group.

Control group: the animals of this group were recruited prospectively from the owners and staff members of the Fundació Hospital Clinic Veterinari of the Universitat Autònoma de Barcelona (FHCV-UAB). The period of the study was September 2013 to July 2015. The group was conformed by healthy cats of more than 1 year of age. Health status was stated on bases of physical exam, biochemistry (calcium, glucose, potassium, total proteins, alanine-amino- transferase (ALT), gamma-glutamyl-transferase (GGT), cholesterol, urea, creatinine), and complete blood count. A SNAP® test to rule out the presence of FIV antibodies and FeLV antigens, and a PCR test to rule out the presence of Bartonella sp were performed.

158 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

Case group: patients with a neoplastic or infectious/inflammatory disease presented at the diagnostic imaging service of the FHCV-UAB, from November 2013 to April 2015, and presented at the diagnostic imaging division of the veterinary faculty of Utrecht University (DDI-UU), from April 2014 to July 2014 were included prospectively. In order to obtain a final diagnosis, the affected lymph nodes were evaluated with cytology (fine needle aspiration (FNA)), histopathology (US-guided biopsy or surgical excision), or at necropsy when possible. Animals in this group were divided in subgroups according to the affected regions as follows: Region 1: head and neck (R1); Region 2: thorax and forelimb (R2); Region 3: abdomen, pelvis, and hindlimb (R3). In each region, the LNs were classified into a neoplastic (N) or inflammatory (I) category according to the following criteria: 1. LNs with a final diagnosis of neoplasia, metastasis, or inflammatory process. 2. LNs with final diagnosis of reactive hyperplasia but with final diagnosis of neoplasia, metastasis, or inflammatory process from the primary lesion. 3. LNs from LCs in the same region with signs of lymphadenopathy (in CT and US) that were not sampled for cytology but with a final diagnosis of neoplasia, metastasis, or inflammatory process from the primary lesion.

Computed Tomography

Control group: Animals were sedated with an intramuscular administration of midazolam14 (0.2mg/kg), burtorphanol15 (0.4mg/kg) and ketamine16 (5mg/kg). 17 Anesthesia was induced with Isoflurane 5% dosage 100% O2 at 4L/min and

maintained with Isoflurane 1.5 – 2% in 100% of O2 at 2L/min. Then patients were positioned on the CT table in dorsal recumbency with the forelimbs and hindlimbs outstretched at the sides. A whole body scan was performed. Acquisitions were done in soft tissue algorithm, before and after the intravenous administration of 600mg/kg of Iopromide18 (300mgI/ml) or Iopamidol19

14 Midazolam 15mg/3ml, Normon, Spain 15 Torbugesic 10 mg/ml, Zoetis, Alcobendas (Madrid), Spain 16 Imalgene 100 mg/ml, Merial, Barcelona, Spain 17 Isoflurane, Abbott Laboratories, Berkshire, UK 18 Ultravist® 300mg/ml, Bayer pharma AG, Berlin, Germany.

159 STUDIES

(300mgI/ml). Scans were performed in a 16 slices helical CT-scanner20. Technical settings were a slice thickness of 0.625 mm, collimation pitch of 1.25mm, 120 kV, 50 - 90 mA, and a matrix of 512x512.

Case group: due to their clinical condition, each patient underwent anesthesia according to the indications of the anesthesia department of each Veterinary Teaching Hospitals. Then patients were positioned in dorsal recumbency. However, the sternal recumbency was used in patients with a critical clinical condition. A regional scan was performed in soft tissue algorithm before and after intravenous administration of 600mg/kg of Iopamidol (300mgI/ml) at the FHCV-UAB, and 600mg/kg of Iobitridol21 (350mgI/ml) at DDI-UU. At the FHCV- UAB, scans were performed with the same CT-scanner and technical settings used for the control group. At DDI-UU, scans were performed in a single-slice helical CT scanner22. Scans were made in helical acquisition mode with slice thickness reconstructions of 0.6 - 1mm, collimation pitch of 1.25mm, 120 kV, 140 – 160 mA, and a matrix of 512x512.

Image analysis: All data were recorded for further analysis using an image archiving and communication system software23. For each lymph node identified, CT characteristics and measurements were performed. Height (short axis) and length (long axis) were obtained to perform a ratio (S/L ratio) in order to compare normal vs abnormal LNs. The height (short axis) was defined as the distance from the ventral to the dorsal border measured in a transverse image. The length (long axis) was determined using a multiplanar reconstruction to generate a sagittal image of the LN at its maximal dimension; an electronic caliper was placed from the rostral/cranial to the caudal border. The shape of the lymph nodes was classified as rounded, elongated, or miscellaneous as previously reported by Beukers et al. (2013) and Nyman, Kristensen, Skovgaard, & McEvoy (2005). Mean attenuation values (Hounsfield units) were determined by placing a circular/oval region of interest (ROI) of 2-4 mm2 at 3 different places at rostral/cranial, middle, and caudal transverse images. In

19 Scanlux® 300mg/ml, Sanochemia pharmazeutika, Neufled/Leitha, Austria. 20 General Electric® Brivo CT 385. 21 Xenetix 350® 350mg/ml, Guerbet, Paris, France. 22 Philips Secura, Philips NV, Eindhoven, the Netherlands. 23 Centricity PACS-IW, GE healthcare (Barcelona) & IMPAX 6, AFGA healthcare (Utrecht)

160 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

small LNs, ROIs were made as large as possible inside the lymph node margins. Measurements were performed on images before and following the administration of contrast medium. The lymph nodes attenuation was compared with the surrounding muscles and was classified as isoattenuating, slightly hypoattenuating, hypoattenuating, hyperattenuating, or heterogeneous. Following the administration of contrast medium the enhancement pattern was classified as homogenous, mildly heterogeneous, heterogeneous, and peripheral enhancement. These parameters were determined following the same criteria as in the author’s previous reports (Tobón Restrepo et al. 2016a, Tobón Restrepo et al. 2016b).

Ultrasonography

Control group: An ultrasound scan was performed to each animal following the

CT scans. Anesthesia was maintained with Isoflurane 1.5 – 2% in 100% of O2 at 2L/min. The hair of some areas near to the localization of superficial lymph nodes was clipped (neck, shoulders, armpits, abdomen, caudal aspect of the stifle joint). The animals were positioned in dorsal recumbency with the neck extended and the forelimbs stretched caudally. Examinations were performed using an Esaote Mylab70 Xvision® machine with a 4 – 13 MHz high frequency linear transducer. Technical settings were adjusted to improve and obtain the optimal images of the LNs in all the animals. Coupling acoustic gel24 was generously applied to ensure an adequate skin-transducer contact. Sagittal and transverse images of each lymph node were recorded.

Case group: Ultrasonography was also performed immediately after the CT. Anesthesia was also maintained according to the protocol chosen by the anesthesia service of either FHCV-UAB or DDI-UU. The hair of the region of interest was clipped as in the control group. Examinations at FHCV-UAB were performed using the same machine as in the control group, but at the DDI-UU were done using a Philips HD11 Ultrasound system® machine with a 7 – 15 MHz high frequency linear transducer.

24 Transonic gel®, Telic, Barcelona, Spain

161 STUDIES

A fine needle aspiration was performed in abnormal lymph nodes of the case group when possible. Images of each lymph node in sagittal and transverse plane were recorded.

Image analysis: For each lymph node, the transducer was placed with the guide pointing rostral/cranial, parallel (or slightly oblique) to the spine, and an image including the largest measurement of the LN (sagittal plane) was recorded. The length was measured using an electronic caliper from the rostral/cranial to the caudal border (long axis). A second measurement was performed in the same image perpendicularly to the length at the thickest point (ventral to dorsal) and was defined as height. These two measurements were used to calculate the S/L ratio of each LN in US. For each lymph node, echogenicity was recorded as hypoechoic, isoechoic, hyperechoic, or heterogeneous when compared to surrounding fat tissue. The shape of each lymph node in ultrasonography was evaluated following the same criteria as in CT. Margins were defined as smooth or irregular.

Statistical analysis

Data of this study were digitalized using Excel (2010)25. The statistical analyses were performed using SPSS software26. The continuous variables obtained in CT and US images in each sub-group were compared using nonparametric statistic as follows: inter-group general comparisons among control, inflammatory and neoplastic LNs were performed with the Kruskal-Wallis test. Inter-group pair comparisons (control and inflammatory; control and neoplastic; inflammatory and neoplastic) were performed with the Mann-Whitney U test. Intra-group pair comparisons (CT and US) were performed with the Wilcoxon test. Fisher exact test was used to evaluate the distributions of the frequencies for the categorical variables. The intra-group and inter-group comparisons were performed for the frequencies of CT and US features. The statistical significance was set to be P<0.05.

25 Microsoft office Excel, 2010 26 IBM SPSS statistic software version 21, 2012.

162 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

Results

Animal description

Control group: 30 cats were recruited in this group. Age and weight average were 3.7 years (range 1.5 – 17) and 4.4 kg (±1.10) respectively. Twenty-nine cats were domestic shorthairs and 1 cat was Persian. The group was conformed by 16.7% entire males (n=5), 20% neutered males (n=6), 30% entire females (n=9) and 33.3% neutered females (n=10). Biochemical determinations and complete blood count were within normal limits. All cats were negative for FIV/FeLV and Bartonella sp. tests.

Case group: 29 cats were recruited in this group. Age and weight averages were 9.1 years (range 0.7 – 16) and 3.6 kg (±1.13) respectively. Twenty-one cats were domestic shorthairs, 5 cats were Persian, and one each of the following breeds: Bombay, British shorthair, and Maine coon. The group was conformed by 17.2% entire males (n=5), 27.6% neutered males (n=8), 31.0% entire females (n=9), 24.1% neutered females (n=7). Eleven cats were presented with complains related to the head and neck and were included in the R1 group. Clinical complains were nasal discharge and stridor (n=4), and one each of the following: intraoral mass, mass in the right masseter region, exophthalmos, head tilt suggestive of vestibular syndrome, FIP suspicion, inspiratory and sometimes expiratory laryngeal/tracheal stridor, and intranasal mass. Eight cats presented with complains related to the thorax and forelimb, were included in the R2 group. Clinical complains comprised mediastinal mass (n=3), dyspnea (n=2), masses in the thoracic wall (n=2), and mass in the left forelimb (n=1). Twelve cats were presented with abdominal complains and were included in the R3 group. Clinical signs were vomiting (acute or chronic) (n=3), hepatic mass (n=2), FIP suspicion (n=2), and one each of the following: intestinal adenocarcinoma, mammary mass, lethargy, severe dermatitis, and weight loss. Of the 29 cats, 2 cats were presented with generalized lymphadenomegaly, therefore they were placed in more than one group.

163 STUDIES

Histopathology results from the FNA and biopsies.

The CT and US images of the control group were analyzed in order to identify all the lymph nodes per lymph center. After the assessment of the CT and US images of the case group, the lymph center from the control group included in the analysis was the one that matched with the affected lymph center in the case group. A total of 723 LNs (R1= 180; R2= 173; R3= 370) from the pre- and postcontrast CT images in the 30 cats of the control group were included. This corresponded to at least one LN per matched lymph center. Additionally, a total of 581 (R1= 180; R2= 85; R3= 316) LNs from the US images were included.

In the case group, a total of 143 and 145 LNs from CT and US respectively were included in the analysis.

In the animals of the R1, 62 (I= 14; N= 48) LNs on CT and 58 (I= 14; N= 44) LNs on US were included, representing at least one LN from the mandibular and retropharyngeal LCs. FNAs were obtained from the mandibular (n=7) and the medial retropharyngeal (n=8) LNs. A biopsy was taken from a mandibular (n=1) LN. Additionally, biopsies results from the nasal mucosa (n=4) and an intraoral mass (n=2) were available. A cytological diagnosis of reactive LN was obtained in 7/8 medial retropharyngeal LN and in 4/7 mandibular LN. A non- diagnostic cytology was obtained in 1/8 medial retropharyngeal LN and in 3/7 mandibular LNs. The biopsy of the mandibular LN concluded a pyogranulomatous lymphadenitis. Four cats with nasal biopsy presented a final diagnosis of chronic rhinitis (n=1), histiocytic sarcoma (n=1), B and T cell lymphoma (n=1), and nasal carcinoma (n=1). One cat with biopsy of an intraoral mass presented a diagnosis of squamous cell carcinoma.

For the R2, 22 (I= 11; N= 11) LNs on CT and 18 (I= 8; N= 10) LNs on US, mainly from the axillary, cranial mediastinal, and ventral thoracic lymph centers, were identified. FNAs were obtained from the sternal (n=3), the cranial mediastinal (n=2), the left axillary (n=1) and the right accessory axillary (n=1) LNs. Additionally, FNA from a mediastinal mass (n=1), a left forelimb mass (n=1), and a biopsy of a mediastinal mass (n=1) and cranial mediastinal LNs (n=2) were available. A cytological diagnosis of reactive LN was obtained in 1/3

164 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

sternal and 1/1 right axillary accessory LNs. A non-diagnostic sample was obtained in 1/3 sternal and 1/1 left axillary LNs. Mediastinal lymphoma was diagnosed in 1/2 cranial mediastinal and 1/3 sternal LNs, and epithelial dysplasia with inflammatory cells was diagnosed in 1/2 cranial mediastinal LNs. Additionally, the FNA of the mediastinal mass was compatible with lymphoma, and the mass in the forelimb was compatible with fibrosarcoma. The biopsies of the mediastinal mass and the cranial mediastinal LN (1/2) were compatible with ectopic thyroid C-cell tumor; mediastinal lymphoma was diagnosed in a cranial mediastinal LN (1/2).

In the R3, 59 (I= 19; N= 40) LNs on CT and 69 (I= 19; N= 50) LNs on US were identified, representing at least one LN from the celiac, cranial and caudal mesenteric LCs. FNAs were obtained from the jejunal (n=5), and the colic (n=2) LNs. Biopsies of the jejunal (n=3), and the caudal mesenteric (n=1) LNs were obtained. FNA of the liver (n=1) and biopsies from the jejunum (n=3), duodenum (n=1), omentum (n=1), and mammary gland (n=1) were available. The cytology was compatible with reactive LN in 3/5 jejunal LNs. A non- diagnostic test was obtained in 1/5 jejunal and 1/2 colic LNs. Lymphoma was diagnosed in 1/5 jejunal LNs, and in 1/2 colic LN. The biopsies of the jejunal LNs were compatible with grade 2 intestinal lymphoma (2/3) and FIP (1/3). Sinus ectasia was diagnosed in the caudal mesenteric LN. The FNA of the liver was suggestive of adenocarcinoma. The biopsies of the jejunum were compatible with grade 2 intestinal lymphoma in 1/3 jejunal samples, and 1/1 duodenum. Intestinal sarcoma was diagnosed in 1/3 jejunal samples. Biopsies of 1/3 jejunal samples and omentum were compatible with feline infectious peritonitis. Mammary carcinoma was diagnosed in 1/1 mammary gland.

CT and US features of LNs.

The mean and standard deviation (SD) of the length, height, S/L ratio of LNs obtained in both techniques for each region are summarized in tables 1 - 3. The qualitative features of the LNs were analyzed by the comparison of the frequencies distribution, and the results are summarized in the table 4.

165 STUDIES

In the R1, the comparisons among groups (Table 1) showed that the LNs were significantly longer in the inflammatory category in both techniques, when compared with the control group, and only significant for the CT, when compared with the neoplastic category. The height of the LNs in CT was higher in the inflammatory categories than in the others. However, there was no statistical difference among groups. On the other hand, the height in US was higher in the inflammatory category, followed by the neoplastic category and the smallest was in the control group. Being this distribution statistically significant in the inter-group comparisons, but not for the comparison between inflammatory and neoplastic. The S/L ratio was higher in the inflammatory category for the CT, but was slightly lower in the US when compared with the control group and the neoplastic categories (the height presented the same value for control and neoplastic). The inter-group comparisons (Table 1) did not show statistical differences for the S/L ratio. The intra-group comparison showed significant differences in height and S/L ratio in both categories of the case group. These were higher in CT when compared with US for both categories. The attenuation of the LNs before contrast administration presented a significantly different distribution between groups. A higher frequency of isoattenuating LNs was identified in all three compared groups, being significantly higher in the inflammatory category. However, the neoplastic category presented 12.5% of the LNs with heterogeneous attenuation, and the overall distribution of attenuation was significantly different when compared with the control group. After contrast administration, more than 50% of the LNs in the three groups showed homogeneous contrast enhancement, being higher in the control group (70.6%). However, both inflammatory and neoplastic LNs presented a significantly higher frequency of heterogeneous contrast enhancement when compared with the control group. The LNs of the inflammatory category presented the highest and the lowest HU before and after contrast administration, respectively. These differences were statistically significant for variables with the control group, but only after contrast with the neoplastic category. Ultrasonographicaly, the LNs of the control group and the inflammatory category where frequently hypoechoic, 82.2 % and 85.7% respectively, differing significantly from the neoplastic category in which heterogeneous echogenicity was found in the 52.3% of the LNs. A hyperechoic

166 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

hilus was only seen in the 6.1% of the LNs in the control group and it was not visualized in the case group. Fusiform LNs were frequently found in the three groups. However, 18.2% of the LNs in the neoplastic group were miscellaneous differing significantly from the 5% in the control group, but not from the 14.3% of the inflammatory category. The margins of the LNs were regular in 100% of the LNs in the control group. Even though regular margins were also frequently found in the case group, irregular margins were present in 14.3% and 6.8% of the inflammatory and neoplastic LNs, respectively. This margins distribution was significantly different between the two case categories and the control group.

In the R2, the assessment of the CT and US images showed that the mean length of the LNs in the inflammatory category was higher when compared with the mean values in the neoplastic category and in the control group, being statistically significant only for the comparison with the control group (Table 2) in both modalities. The neoplastic category presented LNs with higher height and S/L ratio than the other groups. On the US images, the length was significantly higher in the inflammatory category but the height and S/L ratio were significantly higher in the neoplastic category. In the intra-group comparisons (Table 2), the height and S/L ratio differed significantly between CT and US in the inflammatory category. In the neoplastic category, only the height differed significantly between CT and US. The attenuation of the LNs before contrast administration presented a significantly different distribution between the groups as seen in the R1. Isoattenuating LNs were identified frequently in the three groups. However, 20% of the LNs were hyperattenuating in the neoplastic group and this was statistically significant when compared with the control group. After contrast administration, the distribution of the frequencies differed statistically among the groups. The neoplastic and inflammatory categories showed 63.6% and 54.4% of the LNs with heterogeneous contrast enhancement compared with the 0.0% in the control group. The control group showed 67.1% of the LNS with homogeneous contrast enhancement. The mean HU before contrast administration was higher in the neoplastic category and this was significantly different when compared with the control group. The mean HU after the administration of contrast differed significantly between the

167 STUDIES inflammatory and neoplastic categories, being higher in the inflammatory category, but there was no significant difference with the control group. In US, heterogeneous echogenicity was most frequently seen in inflammatory (62.5%) and neoplastic (60.0%) LNs and this was significantly different when compared with the LNs of the control group (41.2% were isoechoic). A hyperechoic hilus was only seen in 12.9% of the LNs in the control group, but was not identified in the case group. Eighty per cent of the neoplastic LNs presented rounded shape differing statistically from the control (55.3% rounded) and the inflammatory (62.5% fusiform) groups. The margins of the LNs were regular in 100% of the LNs in the control group as in the R1. However, the neoplastic and inflammatory LNs presented a 30.0% and 25.0% of LNs with irregular margins, respectively.

In the R3, the inter-group comparisons (Table 3) showed that the mean length and height, in CT and US, were significantly higher in both categories of the case group when compared with the control group. However, the inflammatory category showed larger LNs than the neoplastic category. When the comparison between the two categories of the case group was made, the length and height of the LNs in US were significantly higher in the inflammatory category. The S/L ratio in the case group was statistically higher when compared with the control group. No statistical differences were found between inflammatory and neoplastic S/L ratio of the LNs. In the intra-group comparison (Table 3), only the height in the neoplastic category was statistically higher in the CT compared with US. The attenuation of the LNs before contrast administration showed higher frequencies for the isoattenuating and the slightly hypoattenuating characteristics in all groups. Inflammatory LNs showed a higher frequency of isoattenuating (78.9%) when compared with neoplastic LNs and the control group. After contrast administration, 72.2% of the LNs in the control group showed homogeneous contrast enhancement. The inflammatory and neoplastic LNs showed a significantly higher proportion of heterogeneous contrast enhancement compared with the control group. The mean HU before contrast administration was significantly higher in the LNs of the inflammatory category when compared with the LNs of neoplastic category and the control group. The mean HU after the administration of contrast was significantly lower in the LNs of the neoplastic category when compared with the other groups.

168 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

The control group presented the highest HU after contrast. On US images, the LNs were hypoechoic in 53.5%, 73.7% and 50.0% for the control, inflammatory and neoplastic groups, respectively. Furthermore, the neoplastic LNs presented heterogeneous echogenicity in the 44.0% being only statistically different from the control group. The hyperechoic hilus was only seen in 16.1% of the LNs in the control group but was not identified in the case group. The shape was a very variable characteristic in the control group, with 43.7% of fusiform lymph nodes. On the other hand, the inflammatory category presented 63.2% of miscellaneous LNs, and the neoplastic category 50.0% of rounded LNs, showing significant differences with the control group but not between them. The margins of the LNs were irregular in 52.0% and 42.1% of the neoplastic and inflammatory categories, respectively, being significantly different from the control group that showed regular margins in 94.0% of the LNs.

Discussion

In this study, we show the comparison between the CT and US features of the lymph nodes in a group of healthy cats with the findings in a group of diseased cats. The enrollment of healthy cats (control group) in this study was based in the clinical tests and physical examination. The features of the LNs in the control group concurred with the previous descriptions of normal LNs in cats and dogs. On CT, the LNs were frequently iso- or slightly hypoattenuating in pre-contrast images with homogeneous contrast enhancement (Beukers et al., 2013; Nemanic & Nelson, 2012; Nyman et al., 2005). On US, the LNs were frequently isoechoic or hypoechoic. In some LNs a hyperechoic hilus was visible as previously reported (Nyman et al., 2005; Tobón Restrepo et al., 2015b). The majority of the LNs were elongated with smooth margins. However, rounded and miscellaneous LNs were found in the thorax and abdomen as previously reported (Beukers et al., 2013; D’Anjou, 2008; Hecht & Henry, 2007; Schreurs et al., 2008; Tobón Restrepo, 2015b).

169 STUDIES

Findings in our study suggest that inflammatory LNs are frequently longer and thicker than neoplastic or normal LNs. However, mean values were obtained with the combination of two or more lymph centers per region that could have an influence in this result. Additionally, in the patients with inflammatory diseases, severe pyogranulomatous lymphadenitis was observed, in the context of FIP. Marked LNs enlargement has been described in cats with FIP (Lewis & O’Brien, 2010).

In human medicine, the L/S ratio is used more frequently than the S/L ratio. It has been determined that a L/S ratio of <2 obtained with CT or US is statistically associated with malignancy in head and neck LNs (Mack, Rieger, Baghi, Bisdas, & Vogl, 2008; Mohseni et al., 2014; Steinkamp et al., 1995; Steinkamp, Hosten, Richter, Schedel, & Felix, 1994; Vassallo, Wernecke, Roos, & Peters, 1992). In veterinary medicine, the S/L ratio has been used instead in the differentiation of metastatic vs reactive or normal LNs. (Nyman, Kristensen, Flagstad, & Mcevoy, 2004; Nyman et al., 2005; Nyman & O’Brien, 2007; Tohnosu, Onoda, & Isono, 1989). Nyman et al. (2004) reported a S/L ratio obtained with US of >0.55 to be related with metastatic LNs, and a ratio of <0.55 to be related with a reactive or a normal LN. Likewise, a S/L ratio >0.5 of deep LNs in the dog was significantly associated with neoplasia (De Swarte et al., 2011). Similar results were obtained in our study. The control group showed S/L ratios <0.55 in the three regions, consistent with the previous descriptions. Interestingly, the S/L ratios of the inflammatory and neoplastic categories were <0.55 for both modalities in the R1. We observed that in this region, some mandibular and medial retropharyngeal LNs presented a homogeneous enlargement keeping their elongated shape. This could explain the small variation and the lack of significance in the S/L ratios when comparisons of ratios from normal LN with the ratios from the case group were performed. This situation is very different in the other 2 regions. The neoplastic category showed significantly higher S/L ratio (>0.55) in both modalities for the R2 and R3. This result is consistent with the previous reports that suggest that metastatic LNs presented somehow a rounded shape. However, in the R3, the S/L ratio of the LN in the inflammatory category was >0.55 for the CT images but <0.55 in US images. Previous studies have mentioned the impossibility to ensure an

170 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

accurate measurement of the full length specifically of the jejunal LNs with US (Agthe, Caine, Posch, & Herrtage, 2009). We also found difficult to assess accurately the LNs dimensions of the R3 with US in all the groups. In this region, the LNs of the neoplastic category were significantly bigger with CT when compared with US. An explanation for this is that with CT the assessment of the LN is more accurate due to the lack of superimposition of structures and the possibility to obtain multiplanar reconstructions. Another possible explanation is that the artifacts produced by the gas or feces in the intestines in US can significantly reduce the assessment of the abdominal structures including the LNs. The S/L ratios of the categories from the case group were not significantly different with CT for the R2 and R3 regions. This suggests that inflammatory LNs may increase considerably in size, especially, if the cat is presented with pyogranulomatous lymphadenitis.

In human medicine, LNs are considered malignant in CT images when there is necrotic or cystic changes and/or heterogeneous contrast enhancement, and/or detectable extra nodal tumor spread, and/or they are clearly enlarged in number (Steinkamp et al., 1994; Wunderbaldinger, 2006). In veterinary medicine, mandibular and medial retropharyngeal LNs presenting ellipsoidal shape, heterogeneous contrast enhancement, and hyperattenuating foci (centrally and peripherally) were considered suggestive of lymphadenopathy in cats with nasal polyps (Oliveira, O’Brien, Matheson, & Carrera, 2012). In previous studies, the medial retropharyngeal LNs of cats with nasal lymphoma demonstrated more contrast enhancement and were more homogeneous than those with nasal carcinoma; additionally, the medial retropharyngeal LNs of cats with eosinophilic rhinitis showed a more symmetric and homogeneous pattern of contrast enhancement than cats with suppurative rhinitis (Nemanic et al., 2015). The mandibular and the medial retropharyngeal LNs were most frequently enlarged and showed heterogeneous contrast enhancement with central hypoattenuation in cats with fungal rhinitis and sinusitis (Karnik et al., 2009). Gendler et al. (2008) reported that the mandibular and the medial retropharyngeal LNs were enlarged and showed a rim-enhancement pattern with decrease hilar attenuation in cats with oral squamous cell carcinoma. In our study, the distribution of frequencies for the attenuation before contrast was

171 STUDIES significantly variable. However, a higher frequency of isoattenuating LNs was identified in the three groups for the three regions. There were significant differences in the distribution of attenuation between the control group and the neoplastic category in the three regions. The neoplastic category presented high frequencies of heterogeneous attenuating and hypoattenuating LNs. Meanwhile, the inflammatory category presented a similar frequency distribution of attenuation with the control group. Furthermore, in the R1 and R3, the inflammatory category presented the highest frequency of isoattenuating LNs. The cellular conformation/infiltration of the LN, the amount of cellular degradation, and the alteration in the vasculature could explain the different distribution among groups. We hypothesized that the loss of the perihilar fat could influence in the homogeneity of the attenuation in the case group compared with the normal group. Nevertheless, the exact cause remains unclear. In the postcontrast images, a high frequency of homogeneous enhancement was generally seen in the R1 and R3. In the R2, the case group presented heterogeneous contrast enhancement frequently meanwhile the control group was frequently homogeneous. Our results suggest that inflammatory and neoplastic LNs can exhibit heterogeneous contrast enhancement in similar proportions. A statistical significant difference was only determined in the R2, were all the inflammatory LNs were slightly heterogeneous or showed heterogeneous contrast enhancement, meanwhile neoplastic LNs presented 40% of homogeneous enhancement. Additionally, in the R3, the inflammatory LNs showed peripheral contrast enhancement more frequently than in the neoplastic category and in the control group. We hypothesized that the presence of pyogranulomatous lymphadenitis and sinus ectasia in some LNs of the inflammatory group could have an influence in this result. Elmore, (2006) described that the internal structure of the lymph nodes, with pyogranulomatous changes or sinus ectasia, is replace by intranodal abscesses or cavities filled with lymph. The neoplastic changes in the LNs (due to primary neoplasia or metastasis) normally produce loss of the normal architecture, capsular and perinodal fat invasion, presence of monomorphic population of lymphocytes and changes in nodal size and shape (Elmore, 2006). The LNs of the case group showed significantly higher HU than the control group. After contrast, the HU values were statistically higher in the case

172 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

group for the R1. In the R2 the HU values were similar among groups. In the R3, the HU values in the control group were significantly higher than in the case group. We consider that a higher frequency of heterogeneous contrast enhancement in the case group could influence these results due to the inclusion of areas of reduced contrast enhancement in the ROI.

The internal echogenicity of benign or neoplastic LNs is described in human medicine as isoechoic and hypoechoic in comparison with adjacent tissue, respectively (Khanna, Sharma, Khanna, Kumar, & Shukla, 2011; Mohseni et al., 2014; Steinkamp et al., 1995; Tohnosu et al., 1989). In veterinary medicine, previous reports in dogs referred to heterogeneous echogenicity in LNs in lymphoma and anal sac carcinoma (Llabrés-Díaz, 2004). In a study about abnormal superficial LNs in the dog, normal and reactive LNs were most frequently isoechoic, meanwhile LNs with lymphoma, and metastatic were hypoechoic (Nyman et al., 2005). The results of the present study suggest that neoplastic LNs are usually either hypoechoic or heterogeneous. Meanwhile inflammatory LNs are most frequently hypoechoic. In the R1, the comparison between the categories of the case group showed significant differences suggesting that neoplastic LNs are more heterogeneous with US and inflammatory LNs are most frequently hypoechoic. This result is different to previous reports in which inflammatory LNs were often isoechoic (Nyman et al., 2004; Nyman & O’Brien, 2007), but in agreement with another study that found that inflammatory and neoplastic LNs can have a similar proportion of heterogeneous echogenicity in cats (Kinns & Mai, 2007).

In our results, normal LNs had regular margins and frequently a fusiform shape in the R1 and R3 and were often rounded in the R2 as previously described (Tobón Restrepo, 2015a; Tobón Restrepo et al., 2015b). Other reports mentioned that the US features related with lymphadenopathy are a plump shape with rounded borders and irregular margins (Khanna et al., 2011; Mohseni et al., 2014; Nyman & O’Brien, 2007). In accordance with those, the neoplastic LNs in our study presented a significantly more frequent rounded shape compared with the inflammatory LNs and the control group, that were often fusiform. The LNs in the case group often presented regular margins in the inflammatory category. However, irregular margins were seen in markedly

173 STUDIES enlarged lymph nodes, for example jejunal LNs with pyogranulomatous lymphadenopathy in the cases diagnosed with FIP. The neoplastic category presented the highest frequency of irregular margins in the LNs of the case group.

There are limitations in this study, the prospective nature makes the cases selection challenging. In some cases, it was impossible to repeat the FNA after a laboratory result of non-diagnostic smear due to owner consent or anesthetic risk for the patient. The FNA of LNs has been reported as a technique that is highly operator dependent. Other important factors include that immature or neoplastic lymphoid cells are fragile, and LNs aspirates can yield perinodal fat or other tissues. (Amores-Fuster, Cripps, Graham, Marrington, & Blackwood, 2015). Smears with non-diagnostic results limited the inclusion of more LNs in our study. The repeatability of the diseases was low; even though the study was performed in two university hospitals, the anesthetic risk, the owners’ consent, and the viability of a LN sampling affected the inclusion of many patients. Another limitation in this study is the small number of LNs and the lack of definitive diagnoses from other LNs of the same patient that could be included in the statistical analysis to identify patterns of changes in the different groups of specific diseases.

The results of this study allowed the identification of some features of the lymph nodes that could be suggestive of abnormality. However, a differentiation between inflammatory and neoplastic processes is more challenging. A better assessment and more precise measurements are obtained with CT than with US, especially in the thorax, due to the absence of the artifacts that are present in US. In our study the attenuation of the inflammatory and neoplastic LNs before and after contrast administration could suggest severe changes in the internal structure and vascularization but is not possible to determine a certain degree of malignancy with CT. Ultrasound could have a slight advantage over the CT assessing the vascularization of the LNs with color or power Doppler. Unfortunately, this function was not performed in our study because it is time consuming leading to an undesirable extended anesthesia time. Biopsy and fine needle aspirates (with several repetitions) are still the gold standard to obtain a final diagnosis of lymph nodes pathology.

174 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

In conclusion, LNs of cats with inflammatory or neoplastic processes exhibit features that could be used to differentiate them from normal LNs using both CT and US. Overlapping in the features of both techniques in the categories of the case group prevent an accurate and significant distinction between inflammation and neoplasia. Further studies with a larger sample size could provide more information to complement this study.

References

Agthe, P., Caine, A. R., Posch, B., & Herrtage, M. E. (2009). Ultrasonographic Appearance of Jejunal Lymph Nodes in Dogs Without Clinical Signs of Gastrointestinal Disease. Veterinary Radiology & Ultrasound, 50(2), 195–200.

Amores-Fuster, I., Cripps, P., Graham, P., Marrington, A. M., & Blackwood, L. (2015). The diagnostic utility of lymph node cytology samples in dogs and cats. The Journal of Small Animal Practice, 56(2), 125–9.

Beukers, M., Vilaplana Grosso, F., & Voorhout, G. (2013). Computed Tomographic Characteristics of Presumed Normal Canine Abdominal Lymph Nodes. Veterinary Radiology & Ultrasound, 54(6), 610–617.

D’Anjou, M.-A. (2008). Abdominal cavity, Lymph nodes, and Great Vessels. In D. Penninck & M.-A. D’Anjou (Eds.), Atlas of small animal ultrasonography (First edit., pp. 445 – 463). Ames : Iowa: Blackwell Publishing Ltd.

De Swarte, M., Alexander, K., Rannou, B., D’Anjou, M.-A. A., Blond, L., & Beauchamp, G. (2011). Comparison of sonographic features of benign and neoplastic deep lymph nodes in dogs. Veterinary Radiology and Ultrasound, 52(4), 451–456.

Elmore, S. A. (2006). Histopathology of the Lymph Nodes. Toxicologic Pathology, 34, 425–454.

Gendler, A., Lewis, J. R., Reetz, J. A., & Schwarz, T. Computed tomographic features of oral squamous cell carcinoma in cats: 18 cases (2002–2008). , Journal of the American Veterinary Medical Association 319–325 (2008). - American Veterinary Medical Association.

Hecht, S., & Henry, G. (2007). Sonographic Evaluation of the Normal and Abnormal Pancreas. Ultrasound, 22(3), 115–121.

Henninger, W. (2003). Use of computed tomography in the diseased feline thorax. Journal of Small Animal Practice, 44(2), 56–64.

Karnik, K., Reichle, J. K., Fischetti, A. J., & Goggin, J. M. (2009). Computed Tomographic Findings of Fungal Rhinitis and Sinusitis in Cats. Veterinary Radiology & Ultrasound, 50(1), 65–68.

Khanna, R., Sharma, A. D., Khanna, S., Kumar, M., & Shukla, R. C. (2011). Usefulness of ultrasonography for the evaluation of cervical lymphadenopathy. World Journal of Surgical Oncology, 9(1), 29.

Kinns, J., & Mai, W. (2007). Association between malignancy and sonographic heterogeneity in canine and feline abdominal lymph nodes. Veterinary Radiology & Ultrasound, 48(6), 565– 569.

175 STUDIES

Lewis, K. M., & O’Brien, R. T. (2010). Abdominal ultrasonographic findings associated with feline infectious peritonitis: A retrospective review of 16 cases. Journal of Animal Hospital Association, 46, 152–160.

Li, L., Mori, S., Kodama, M., Sakamoto, M., Takahashi, S., & Kodama, T. (2013). Enhanced sonographic imaging to diagnose lymph node metastasis: importance of blood vessel volume and density. Cancer Research, 73(7), 2082–92.

Llabrés-Díaz, F. J. (2004). Ultrasonography of the medial iliac lymph nodes in the dog. Veterinary Radiology & Ultrasound, 45(2), 156–165.

Mack, M. G., Rieger, J., Baghi, M., Bisdas, S., & Vogl, T. J. (2008). Cervical lymph nodes. European Journal of Radiology, 66(3), 493–500.

Mohseni, S., Shojaiefard, A., Khorgami, Z., Alinejad, S., Ghorbani, A., & Ghafouri, A. (2014). Peripheral lymphadenopathy: Approach and diagnostic tools. Iranian Journal of Medical Sciences, 39(2 SUPPL.), 158–170.

Nemanic, S., Hollars, K., Nelson, N. C., & Bobe, G. (2015). Combination of Computed Tomographic Imaging Characteristics of Medial Retropharyngeal Lymph Nodes and Nasal Passages Aids Discrimination Between Rhinitis and Neoplasia in Cats. Veterinary Radiology & Ultrasound, 56(6), 617–627.

Nyman, H. T., Kristensen, A. T., Flagstad, A., & Mcevoy, F. J. (2004). A review of the sonographic assessment of tumor metastases in liver and superficial lymph nodes. Veterinary Radiology & Ultrasound, 45(5), 438–448.

Nyman, H. T., Kristensen, A. T., Skovgaard, I. M., & McEvoy, F. J. (2005). Characterization of normal and abnormal canine superficial lymph nodes using gray-scale B-mode, color flow mapping, power, and spectral doppler ultrasonography: A multivariate study. Veterinary Radiology and Ultrasound, 46(5), 404–410.

Nyman, H. T., & O’Brien, R. T. (2007). The Sonographic Evaluation of Lymph Nodes. Clin Tech Small Anim Pract, 22(3), 128–137.

Oliveira, C. R., O’Brien, R. T., Matheson, J. S., & Carrera, I. (2012). Computed tomographic features of feline nasopharyngeal polyps. Veterinary Radiology & Ultrasound, 53(4), 406– 11.

Salwei, R. M., O’Brien, R. T., & Matheson, J. S. (2005). Characterization of lymphomatous lymph nodes in dogs using contrast harmonic and power doppler ultrasound. Veterinary Radiology and Ultrasound, 46(5), 411–416.

Schreurs, E., Vermote, K., Barberet, V., Daminet, S., Rudorf, H., & Saunders, J. H. (2008). Ultrasonographic Anatomy of Abdominal Lymph Nodes in the Normal Cat. Veterinary Radiology & Ultrasound, 49(1), 68–72.

Steinkamp, H. J., Cornehl, M., Hosten, N., Pegios, W., Vogl, T., & Felix, R. (1995). Cervical lymphadenopathy: Ratio of long- to short-axis diameter as a predictor of malignancy. British Journal of Radiology, 68(807), 266–270.

Steinkamp, H. J., Hosten, N., Richter, C., Schedel, H., & Felix, R. (1994). Enlarged cervical lymph nodes at helical CT. Radiology, 191, 795–798.

Tobón Restrepo, M. (2015a). Ultrasound of the abdominal cavity, lymph nodes and large vessels. In R. Novellas Torroja, E. Dominguez Miño, Y. Espada Gerlach, Y. Martínez Pereira, & M. Tobón Restrepo (Eds.), Diagnostic ultrasound in cats (1st ed., pp. 211 – 229). Zaragoza: Spain: Servet editorial - Grupo Asís Biomedia S.L.

Tobón Restrepo, M., Novellas, R., Espada, Y., Aguilar, A., Moll, X., & Boroffka, S. (2015b). Ultrasonography and computed tomography features of lymph nodes in healthy cats. In Abstracts from the 2014 european veterinary diagnostic imaging annual conference.

176 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

Veterinary Radiology & Ultrasound (Vol. 56, p. 700). Utrecht, The Netherlands.

Tobón Restrepo, M., Espada, Y., Aguilar, A., Moll, X., & Novellas, R. (2016a). Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part I. The head, neck, thorax, and forelimb. Pending of publication.

Tobón Restrepo, M., Novellas, R., Aguilar, A., Moll, X., & Espada, Y. (2016b). Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats: Part II. The abdomen and hindlimb. Pending of publication.

Tohnosu, N., Onoda, S., & Isono, K. (1989). Ultrasonographic evaluation of cervical lymph node metastases in esophageal cancer with special reference to the relationship between the short to long axis ratio (S/L) and the cancer content. JCU J Clin Ultrasound, 17(2), 101– 106.

Vassallo, P., Wernecke, K., Roos, N., & Peters, P. (1992). Differentiation of benign from malignant superficial lymphadenopathy: the rol of high-resolution US. Radiology, (183), 215–220.

Wunderbaldinger, P. (2006). Problems and prospects of modern lymph node imaging. Modern Lymphnode Imaging, 58(3), 325–337.

Table 1. Mean and SD for the quantitative variables obtained with CT and US of the LNs in the R1 (head and neck). Inter and intra group comparison among techniques.

Features Control (C) Inflammatory (I) Neoplastic (N) Computed tomography n Mean SD n Mean SD n Mean SD 180 14.38 5.59 14 16.74 4.62 48 13.64 5.64 Length (mm) 180 6.04 4.49 14 8.36 6.24 48 5.77 4.21 Height (mm) S/L ratio 180 0.39 0.20 14 0.50 0.41 48 0.44 0.41

HU PC 180 39.76 11.28 14 49.50 8.27 48 43.67 16.17

HU AC 180 132.94 30.93 14 117.74 5.74 48 140.34 21.08

Ultrasonography

180 11.01 3.59 14 15.31 3.49 44 13.14 5.42 Length (mm) 180 3.32 1.25 14 4.35 2.02 44 4.06 2.01 Height (mm) S/L ratio 180 0.31 0.09 14 0.28 0.07 44 0.31 0.09

P-Value Features Comparisons CT US Inflammatory Neoplastic Length C x I x N* 0.068 0.000 C x I† 0.044 0.000 C x N† 0.356 0.009 I x N† 0.025 0.050 CT x US 0.184 0.058

Height C x I x N* 0.241 0.018 C x I† 0.099 0.028 C x N† 0.906 0.037 I x N† 0.113 0.636 CT x US 0.013 0.000

S/L ratio C x I x N* 0.964 0.398 C x I† 0.791 0.182 C x N† 0.975 0.902 I x N† 0.801 0.217 CT x US 0.041 0.003

HU PC C x I x N* 0.014 C x I† 0.002 C x N† 0.503 I x N† 0.080 HU AC C x I x N* 0.009 C x I† 0.047 C x N† 0.109

I x N† 0.000 * Kruskal-Wallis test, † Mann-Whitney U test, Wilcoxon test. Statistic significant when P>0.05 . S/L ratio: short / long axis ratio; HU PC: Hounsfield units precontrast; HU AC: Hounsfield units after contrast.

Table 2. Mean and SD for the quantitative variables obtained with CT and US of the LNs in the R2 (thorax and forelimb). Inter and intra group comparison among techniques.

Features Control (C) Inflammatory (I) Neoplastic (N) Computed tomography n Mean SD n Mean SD n Mean SD Length (mm) 173 10.89 4.65 11 17.28 4.74 11 12.53 7.86 Height (mm) 173 4.21 1.67 11 8.86 2.77 11 9.09 6.51 S/L ratio 173 0.43 0.18 11 0.53 0.16 11 0.83 0.56 HU PC 173 20.81 23.56 11 35.00 12.94 10 44.40 15.56 HU AC 173 88.95 34.06 11 99.42 21.16 11 82.24 11.38 Ultrasonography

Length (mm) 85 7.65 2.35 8 12.08 4.88 10 11.72 8.53 Height (mm) 85 3.37 1.12 8 3.99 0.91 10 6.92 4.59 S/L ratio 85 0.46 0.15 8 0.36 0.13 10 0.65 0.23

P-Value Features Comparisons CT US Inflammatory Neoplastic Length C x I x N* 0.026 0.001 C x I† 0.008 0.000 C x N† 0.891 0.267 I x N† 0.071 0.657 CT x US 0.069 0.594

Height C x I x N* 0.000 0.026 C x I† 0.000 0.074 C x N† 0.018 0.030 I x N† 0.818 0.327 CT x US 0.017 0.017

S/L ratio C x I x N* 0.017 0.008 C x I† 0.053 0.056 C x N† 0.027 0.018 I x N† 0.577 0.008 CT x US 0.036 0.374

HU PC C x I x N* 0.001 C x I† 0.054 C x N† 0.001 I x N† 0.159 HU AC C x I x N* 0.201 C x I† 0.206 C x N† 0.252

I x N† 0.045 * Kruskal-Wallis test, † Mann-Whitney U test, Wilcoxon test. Statistic significant when P>0.05. S/L ratio: short / long axis ratio; HU PC: Hounsfield units precontrast; HU AC: Hounsfield units after contrast.

Table 3. Mean and SD for the quantitative variables obtained with CT and US of the LNs in the R3 (abdomen, pelvis and hindlimb). Inter and intra group comparison among techniques.

Features Control (C) Inflammatory (I) Neoplastic (N) Computed tomography n Mean SD n Mean SD n Mean SD Length (mm) 370 11.56 9.05 19 20.36 13.50 40 13.61 7.94 Height (mm) 370 4.18 2.02 19 8.74 3.69 40 7.62 3.95 S/L ratio 370 0.50 0.30 19 0.63 0.47 40 0.64 0.34 HU PC 370 32.72 19.37 19 42.33 8.80 40 35.18 14.40 HU AC 370 111.63 31.71 19 106.91 23.37 40 103.71 22.73 Ultrasonography

Length (mm) 316 9.24 6.33 19 16.89 9.92 50 11.69 7.06 Height (mm) 316 3.20 1.18 19 7.68 4.01 50 5.55 2.95 S/L ratio 316 0.42 0.18 19 0.51 0.19 50 0.54 0.19

P-Value Features Comparisons CT US Inflammatory Neoplastic Length C x I x N* 0.000 0.000

C x I† 0.000 0.001 C x N† 0.006 0.020 I x N† 0.085 0.039 CT x US 0.619 0.722

Height C x I x N* 0.000 0.000 C x I† 0.000 0.000 C x N† 0.000 0.000 I x N† 0.311 0.026 CT x US 0.381 0.014

S/L ratio C x I x N* 0.007 0.000 C x I† 0.321 0.037 C x N† 0.002 0.000 I x N† 0.408 0.510 CT x US 0.246 0.092

HU PC C x I x N* 0.032 C x I† 0.009 C x N† 0.682 I x N† 0.032 HU AC C x I x N* 0.085 C x I† 0.461 C x N† 0.033

I x N† 0.372 * Kruskal-Wallis test, † Mann-Whitney U test, Wilcoxon test. Statistic significant when P>0.05. S/L ratio: short / long axis ratio; HU PC: Hounsfield units precontrast; HU AC: Hounsfield units after contrast.

Table 4. Comparison of CT and US characteristics of LNs in the control and case groups per regions.

R1 R2 R3 Features P- P- P- Comparisons Comparisons Comparisons Control Inflammatory Neoplastic Value Control Inflammatory Neoplastic Value Control Inflammatory Neoplastic Value Computed tomography n % n % n % n % n % n % n % n % n % Precontrast Isoattenuating 91 50.6 11 78.6 25 52.1 C x I x N 0.001 80 46.2 5 45.5 4 40.0 C x I x N 0.013 172 46.5 15 78.9 21 52.5 C x I x N 0.000 Slightly hypoattenuating 61 33.9 2 14.3 11 22.9 C x I 0.156 56 32.4 2 18.2 1 10.0 C x I 0.133 168 45.4 3 15.8 10 25.0 C x I 0.025 Hypoattenuating 28 15.6 1 7.1 6 12.5 C x N 0.000 4 2.3 2 18.2 1 10.0 C x N 0.011 5 1.4 1 5.3 7 17.5 C x N 0.000 Hyperattenuating 0 0.0 0 0.0 0 0.0 I x N 0.410 2 1.2 0 0.0 2 20.0 I x N 0.756 3 0.8 0 0.0 0 0.0 I x N 0.310 Heterogeneous 0 0.0 0 0.0 6 12.5 31 17.9 2 18.2 2 20.0 22 5.9 0 0.0 2 5.0 After contrast

Homogeneous CH 127 70.6 7 50.0 26 54.2 C x I x N 0.000 116 67.1 0 0.0 4 36.4 C x I x N 0.000 269 72.7 7 36.8 13 32.5 C x I x N 0.000 Slightly heterogeneous CH 43 23.9 1 7.1 4 8.3 C x I 0.000 20 11.6 5 45.5 0 0.0 C x I 0.000 52 14.1 3 15.8 13 32.5 C x I 0.001 Heterogeneous CH 10 5.6 6 42.9 18 37.5 C x N 0.000 0 0.0 6 54.4 7 63.6 C x N 0.000 18 4.9 5 26.3 10 25.0 C x N 0.000 Peripheral enhancement 0 0.0 0 0.0 0 0.0 I x N 0.899 37 21.4 0 0.0 0 0.0 I x N 0.012 31 8.4 4 21.1 4 10.0 I x N 0.445 Ultrasonography

Isoechoic 22 12.2 2 14.3 5 11.4 C x I x N 0.000 35 41.2 0 0.0 0 0.0 C x I x N 0.001 82 25.9 2 10.5 2 4.0 C x I x N 0.000 Hypoechoic 148 82.2 12 85.7 16 36.4 C x I 1.000 22 25.9 3 37.5 4 40.0 C x I 0.010 169 53.5 14 73.7 25 50.0 C x I 0.352 Hyperechoic 0 0.0 0 0.0 0 0.0 C x N 0.000 11 12.9 0 0.0 0 0.0 C x N 0.003 7 2.2 0 0.0 1 2.0 C x N 0.000 Heterogeneous 10 5.6 0 0.0 23 52.3 I x N 0.000 17 20.0 5 62.5 6 60.0 I x N 1.000 58 18.4 3 15.8 22 44.0 I x N 0.074 Shape

Rounded 1 0.6 0 0.0 2 4.5 C x I x N 0.005 47 55.3 2 25.0 8 80.0 C x I x N 0.014 89 28.2 7 36.8 25 50.0 C x I x N 0.000 Elongated 170 94.4 12 85.7 34 77.3 C x I 0.243 37 43.5 5 62.5 1 10.0 C x I 0.064 138 43.7 0 0.0 3 6.0 C x I 0.000 Miscellaneous 9 5.0 2 14.3 8 18.2 C x N 0.002 1 1.2 1 12.5 1 10.0 C x N 0.036 89 28.2 12 63.2 22 44.0 C x N 0.000 I x N 1.000 I x N 0.027 I x N 0.363 Margins C x I x N 0.000 C x I x N 0.000 C x I x N 0.000

Regular 180 100.0 12 85.7 41 93.2 C x I 0.005 85 100.0 6 75.0 7 70.0 C x I 0.007 297 94.0 11 57.9 24 48.0 C x I 0.000 Irregular 0 0.0 2 14.3 3 6.8 C x N 0.007 0 0.0 2 25.0 3 30.0 C x N 0.001 19 6.0 8 42.1 26 52.0 C x N 0.000 I x N 0.585 I x N 1.000 I x N 0.592 Statistic significant when P>0.05

STUDIES 181

Figure 1. Lymph nodes in three patients of the R1: (a – c). Normal appearance of the retropharyngeal LN (RPLN). The LN (yellow arrow) is visible in CT transverse images with a fusiform shape, isoattenuating to the surrounding musculature in the precontrast image (a), and shows homogeneous contrast enhancement (b). The occipital bone (Oc) and the external jugular vein (open arrow head) are indicated. On the US image (c) from the same patient, a heterogeneous (hypoechoic) LN with fusiform shape is visible (between cursors). The carotid artery (asterisk) is indicated.

(d – f). Inflammatory RPLN in a patient with pyogranulomatous lymphadenitis. The LNs (yellow arrow) in CT transverse images are asymmetrically enlarged, with a miscellaneous shape, slightly hypoattenuating compared with the surrounding muscle in precontrast (d) and with slightly heterogeneous contrast enhancement (e). The tympanic bullae (B) and external jugular vein (open arrow head) are indicated. On US images (f) of the same patient, the LN is enlarged, mainly hypoechoic with mild irregular margins and fusiform shape (between cursors). The carotid artery (asterisk) is partially seen.

(g – i). Neoplastic RPLN in a patient with oral squamous cell carcinoma. The LNs (yellow arrow) in CT transverse images are asymmetrically enlarged with fusiform shape, slightly hypoattenuating (g) compared to the surrounding muscles and with homogeneous (left) and slightly heterogeneous (right) contrast enhancement (h). The atlas and dens of the axis (A) and the jugular vein (open arrow head) are indicated. On US images (i) of the same patient, a heterogeneous LN with fusiform shape and slightly irregular margins is visible (between cursors). The carotid artery (asterisk) is indicated.

182 Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

Figure 2. Lymph nodes in three patients of the R2: (a – c). Normal appearance of the sternal LN (SLN). The LN (yellow arrow) is visible in CT sagittal images with a fusiform shape, isoattenuating to the musculature with a hypoattenuating center in the precontrast image (a), and with heterogeneous contrast enhancement (b). The heart (H) and the 3rd sternebra (S) are indicated. US image (c) of the same patient, in which a hypoechoic LN with a hyperechoic central line (hilus) and fusiform shape is visible (between cursors). The 3rd rib (blue arrow) is indicated.

(d – f). Inflammatory accessory axillary LNs in a patient with multiple abscesses in the thoracolumbar region. The LNs (yellow arrow) in CT transverse images are asymmetrically enlarged, with a rounded shape, isoattenuating to the musculature in precontrast images (d) and show homogeneous contrast enhancement (e). The heart (H), 6th thoracic vertebra (Th6), and the muscles latissimus dorsi (1), serratus ventralis (2), and pectoralis (3) are indicated. On US images (f) of the same patient, the LN (yellow arrow) is enlarged, hypoechoic, with irregular margins and a miscellaneous shape. The 5th rib (R) is indicated.

(g – i). Neoplastic SLN in a patient with mediastinal lymphoma. The LN (yellow arrow) in CT sagittal images is enlarged, has a miscellaneous shape, and heterogeneous attenuation in pre- (g) and postcontrast (h) images. The 3rd sternebra (S), the heart (H) and pleural effusion (E) are indicated. On US images (i) of the same patient, a heterogeneous (mainly isoechoic) LN with miscellaneous shape and irregular margins is visible (between cursors). The cartilage of the 4th rib (asterisk) is indicated.

STUDIES 183

Figure 3. Lymph nodes in three patients of the R3: (a – c). Normal appearance of the jejunal LN (JLN). The LN (yellow arrow) is visible in CT dorsal images with a miscellaneous shape, isoattenuating to the muscles in precontrast image (a), and with homogeneous contrast enhancement (b). The spleen (S) and the ilecocolic junction (I) are indicated. US image (c) of the same patient in which a hypoechoic LN with miscellaneous shape is visible (yellow arrows). The spleen (S) is indicated.

(d – f). Inflammatory JLNs in a patient with pyogranulomatous lymphadenitis associated with feline infectious peritonitis (FIP). The LNs (asterisks) in CT dorsal images are symmetrically enlarged, with a miscellaneous shape, isoattenuating to the muscles in precontrast (d) and with heterogeneous contrast enhancement (e). The spleen (S), and ileocolic junction (I) are indicated. US images (f) of the same patient; the LN (between cursors) is enlarged, shows heterogeneous echogenicity with irregular margins and miscellaneous shape.

(g – i). Neoplastic JLNs in a patient with alimentary lymphoma. The LNs (yellow arrow) in CT dorsal images are enlarged, with a miscellaneous shape, isoattenuating in precontrast images (g) and with homogeneous contrast enhancement (h). The spleen (S) and the colon (C) are indicated. US images (i) of the same patient; heterogeneous LNs with miscellaneous shape and irregular margins are visible (between cursors). The jejunal vessels (asterisk) are partially seen between the 2 JLNs.

5. GENERAL DISCUSSION

GENERAL DISCUSSION 187

A better understanding in the assessment of cat’s lymph nodes with CT and US is presented in this thesis. There are 19 lymph centers in the cat according to previous reports (NAV, 2012; Saar & Getty, 1982; Tompkins, 1993). In this thesis, at least a representative lymph node from 17 lymph centers was successfully identified. The dorsal thoracic and iliofemoral lymph centers could not be identified in any cat and with any of the methods used in this study.

Anatomic, computed tomographic, and ultrasonographic assessment of the lymph nodes in healthy adult cats

The dissection of 6 cadavers allowed the identification of at least one lymph node per lymph center and provided the reference landmarks to search the same lymph nodes on CT and US images. As previously reported, the lymph nodes are embedded in fat tissue (Saar & Getty, 1982) and the amount of this is proportional to the body condition of the animal. The length, height, and width of each lymph node were obtained and compared among techniques (anatomy, CT, and US). The measurements obtained in the anatomic study were in general shorter than those obtained with CT and US. Additionally, these measurements were also shorter than those previously reported in anatomic studies in cats (Saar & Getty, 1982; Sugimura, Kudo, & Takahata, 1955, 1956, 1958, 1959). We hypothesized that the age of the included animals (very young animal were included in previous studies) and dissection process made after 24 hours of dead (unknown degree of dehydration or volume loss after the blood supply has stop) could have an influence in these results.

The number of LN identified with CT per lymph center was higher than with the other techniques. The higher number of identified LNs was influenced by the lack of superimposition of structures in CT, the body condition of the animals, and the use of contrast medium. The measurements obtained with CT were in general higher than those obtained in anatomy and with US. Additionally, the comparisons showed that there were statistically significant

188 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

differences in the measurements, specially the height, among techniques. These significant differences were also more frequently identified in the head and neck lymph centers. We found that the possibility of obtaining reconstructed images (multiplanar reconstruction) with the CT allowed a much better understanding and assessment of the lymph nodes. In our study, CT images were performed with a slice thickness of 0.6mm, which produces high resolution reconstructions with a very clear contour of the structures. We found that with reconstructed images, the overestimation of the length and height is prevented in comparison with the transverse slices. This is because a complete parallel position of the LN with the table cannot be ensured, as previously described (Sarah Nemanic & Nelson, 2012). Also, in thin animals in which the lymph nodes showed border effacement with the surrounding tissue, the use of contrast medium improved their identification. The LNs identified with CT were most frequently isoattenuating or slightly hypoattenuating to the surrounding musculature, with homogeneous contrast enhancement. Similar findings have been previously described in humans, dogs, and for some LNs in cats (Beukers, Vilaplana Grosso, & Voorhout, 2013; S Nemanic & Nelson, 2012; Wunderbaldinger, 2006). However, some LNs (e.g. axillary, sternal, and popliteal LNs) presented a isoattenuating periphery with a hypoattenuating center (before and after contrast administration) due to a fatty hilus (Beukers et al., 2013; Kneissl & Probst, 2007; Rossi, Patsikas, & Wisner, 2011).

The number of identified LNs with US was reduced compared with the other techniques. This result was directly influenced by the superimposition of the gastrointestinal tract with gas or feces that produced artifacts in the images; the impossibility to assess the thoracic cavity for the mediastinal, dorsal thoracic, and bronchial lymph centers; difficulty in identifying some LNs in animals with an elevated body condition; and, in a lesser degree, the poor skin-probe contact in some areas that were not completely clipped and the experience of the operator for the initial scans. The measurements obtained with US were significantly different from the CT, specially in those LNs were the position in transverse images on CT were somewhat oblique, and in

GENERAL DISCUSSION 189

those were artifacts produced in US resulted in underestimation of their size. The LNs measurement with US in our study showed some differences with those reported by Schreur et al. (2008) for the abdomen. The cause for these differences remains unclear, it is possible that the sample size, fasting period of the cats, the scanning planes, and the interobserver variability were contributing factors.

In US, most LNs were isoechoic or hypoechoic to the surrounding fat tissue. The majority of LNs were fusiform or rounded. Nevertheless, the MRLNs, the dorsal SCLNs, and the accessory axillary LNs (AAxLNs) had miscellaneous shape. Some LNs presented a hypoechoic / isoechoic periphery with a hyperechoic center when compared with the musculature (e.g. axillary, sternal, and popliteal) due to the presence of a fatty hilus. This description is similar to previous studies (Agthe, Caine, Posch, & Herrtage, 2009; D’Anjou, 2008; S Nemanic & Nelson, 2012; Nyman, Kristensen, Skovgaard, & McEvoy, 2005; Nyman & O’Brien, 2007).

Assessment of normal and abnormal lymph nodes in cats using computed tomography and ultrasonography

Comparisons of the LNs features from the head, neck, forelimbs, thorax, abdomen, and hindlimbs between healthy cats and cats with inflammatory and neoplastic diseases were performed.

The quantitative variables, length, height and S/L ratio, obtained with both techniques, and the HU before contrast administration, were significantly different between the control and the case group for the region of the thorax and forelimb, and the region of the abdomen and hindlimb. In the region of the head and neck, statistical differences between the control and case groups were only observed for the HU before and after contrast administration and for the length and height obtained with US. The S/L ratios were no significant different between the inflammatory and neoplastic categories in both techniques for the region of the abdomen and hindlimb, and only for the CT in the region of the thorax and forelimb. It was observed

190 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

that inflammatory LNs may increase considerable in size, especially, if a pyogranulomatous process is taking place which could explain the similar ratios and the lack of statistical difference.

The LNs of the control group on CT images were frequently iso- or slightly hypoattenuating in precontrast images and showed homogeneous contrast enhancement (Beukers et al., 2013; Kneissl & Probst, 2007; Rossi et al., 2011). In the case group, the neoplastic category presented high frequencies of heterogeneous attenuating and hypoattenuating LNs. Meanwhile, the inflammatory category presented a similar frequency of attenuations to the control group. Furthermore, in the region of the thorax and forelimb and the region of the abdomen and hindlimb, the inflammatory category presented the highest frequency of isoattenuating LNs. The heterogeneous and hypoattenuating appearance of the most neoplastic and some inflammatory LNs before and after contrast administration could suggest severe changes in the internal structure and vascularization but a certain degree of malignancy was not possible to determine with CT.

On US, LNs in the control group were frequently isoechoic or hypoechoic and some of them presented a hyperechoic hilus, characteristics described in previous studies (Agthe et al., 2009; D’Anjou, 2008; S Nemanic & Nelson, 2012; Nyman et al., 2005; Nyman & O’Brien, 2007). In the case group, neoplastic LNs were either hypoechoic or heterogeneous, meanwhile inflammatory LNs were more frequently hypoechoic. In the region of the head and neck, results suggest that neoplastic LNs are more heterogeneous and inflammatory LNs are more frequently hypoechoic. This result is different to previous reports in which inflammatory LNs were often isoechoic (Nyman, Kristensen, Flagstad, & Mcevoy, 2004; Nyman & O’Brien, 2007), but in agreement with another study that found that inflammatory and neoplastic LNs can have a similar proportion of heterogeneous echogenicity in cats (Kinns & Mai, 2007).

The majority of the LNs in the control group were elongated (a small proportion presented rounded and miscellaneous shape specially in the thorax and abdomen) with smooth margins as already described (D’Anjou,

GENERAL DISCUSSION 191

2008; Nyman et al., 2004; Nyman & O’Brien, 2007). As previously described, the neoplastic LNs in our study presented a rounded shape more frequently than the inflammatory LNs and the control group, that were often fusiform (Khanna, Sharma, Khanna, Kumar, & Shukla, 2011; Mohseni et al., 2014; Nyman & O’Brien, 2007). The LN in the inflammatory category often presented regular margins. However, irregular margins were seen in markedly enlarged lymph nodes (e.g. jejunal LNs with pyogranulomatous lymphadenitis). The neoplastic category presented the highest frequency of irregular margins in the LNs of the case group.

Further studies with a larger sample size and comparing specific neoplastic and inflammatory process are needed for a better understanding of the changes in the LNs in the disease cat. This thesis could be the beginning of future research lines involving other images techniques and description of different features of the feline LNs (that those described in this thesis) that are of common use in human medicine.

6. CONCLUSIONS

CONCLUSIONS 195

1. The LNs of healthy cats can be successfully identified with CT images, and the use of contrast medium and multiplanar reconstruction improves the assessment of their size. 2. The identification of LNs in healthy cats with US could be achieved for almost all corporal regions but the thoracic cavity (especially dorsal thoracic, mediastinal, and bronchial lymph centers). The presence of air in the lung and intestines, food in the stomach and fecal material in the colon, are factors that limit the assessment. 3. The LNs of healthy cats on CT images were frequently iso- or slightly hypoattenuating in precontrast images with homogeneous contrast enhancement, and on US were frequently isoechoic or hypoechoic and a low percentage presented a hyperechoic hilus. 4. The measurements on CT were significantly larger when compared with US and anatomy. The measurements in both techniques are proposed as reference values. 5. The short-to-long-axis (S/L) ratio of the LNs obtained with US and CT in diseased cats was frequently significantly different from the healthy cat. 6. The neoplastic category presented high frequencies of heterogeneous attenuating and hypoattenuating LNs on CT and were frequently hypoechoic or heterogeneous on US. Meanwhile, the inflammatory category presented higher frequencies of iso- to slightly hypoattenuating appearance on CT and were more frequently hypoechoic. 7. Overlapping in the features of both techniques in the categories of the case group prevents an accurate distinction between inflammation and neoplasia.

7. SUMMARY

SUMMARY 199

The lymph nodes (LNs) are structures that play an important role in the diagnosis and prognosis of neoplastic and infectious diseases. Descriptions about the changes in size (increase of a short-to-long-axis ratio (S/L)>0.5), the shape (from elongated to rounded) and the internal structure of LNs assessed with ultrasound are associated with lymphadenopathy in humans and dogs. However, there are few studies assessing these parameters in cats. The aims of the studies described in this thesis were (i) to assess the ability to identify the LNs of healthy cats with computed tomography (CT) and ultrasonography (US) and to compare the imaging measurements with measurements from normal anatomic values; (ii) to characterize the LNs in diseased cats using CT and US; and (iii) to assess the ability of each imaging technique (CT and US) to discriminate between neoplastic and inflammatory changes in the LNs. The number of LNs identified with CT was higher than with US and anatomy. The measurements obtained with CT were in general higher than those obtained in anatomy and with US. Additionally, the comparisons showed that there were statistically significant differences in the measurements, specially the height, among techniques. With CT-reconstructed images, the overestimation of the length and height is prevented in comparison with the transverse slices. The LNs identified with CT showed a similar appearance than previously described being most frequently isoattenuating or slightly hypoattenuating to the surrounding musculature, with homogeneous contrast enhancement. Most LNs were isoechoic or hypoechoic to the surrounding fat tissue on US images. The majority of LNs were fusiform or rounded. Some LNs presented a hypoechoic/isoechoic periphery with a hyperechoic center when compared with the musculature, due to the presence of a fatty hilus. The length, height, and S/L ratio, obtained with CT and US, and the HU before contrast administration were significantly different between the control and the case group for the thorax and forelimb, and the abdomen and hindlimb regions. In the region of the head and neck, statistical differences between the control and case groups were only observed for the HU before and after contrast administration and for the length and height obtained with

200 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

US. The S/L ratios were no significant different between the inflammatory and neoplastic categories in both techniques for the region of the abdomen and hindlimb, and only for the CT in the region of the thorax and forelimb. The neoplastic category presented high frequencies of heterogeneous attenuating and hypoattenuating LNs. Meanwhile, the inflammatory category presented a similar frequency of attenuations to the control group. The heterogeneous and hypoattenuating appearance of the most neoplastic and some inflammatory LNs before and after contrast administration could suggest severe changes in the internal structure and vascularization but a certain degree of malignancy was not possible to determine with CT. On US, the neoplastic LNs were most frequently hypoechoic or heterogeneous, meanwhile inflammatory LNs were more frequently hypoechoic. In the region of the head and neck, results suggest that neoplastic LNs are more heterogeneous and inflammatory are more frequently hypoechoic. The neoplastic LNs in our study presented a rounded shape more frequently than inflammatory LNs and the control group, in which they were often fusiform. The inflammatory LNs often presented regular margins. However, irregular margins were seen in markedly enlarged lymph nodes, especially in neoplastic LNs. In conclusion, identification of the LNs of the cat is possible with both techniques. Computed tomography was more accurate in the detection and performance of measurements of the LNs than US. However, overlapping in the features of both techniques in the categories of the case group prevents an accurate distinction between inflammation and neoplasia.

8. RESUM

RESUM 203

Els nòduls limfàtics (NLs) són estructures importants en el diagnòstic i pronòstic de malalties tant neoplàsiques com infeccioses. Estudis previs han associat els canvis observats per ecografia respecte a la grandària (augment del ratio eix-curt-eix llarg (S/L) a > 0,5), la forma (de allargada a arrodonida) i l'estructura interna dels NLs en éssers humans i gossos amb limfadenopatia. No obstant això, hi ha molt pocs estudis en gats en els quals es descriguin aquests paràmetres amb tomografia computada (TC) o ecografia (US). Els objectius d’aquesta tesi van ser: (i) avaluar la capacitat de la TC i US per identificar els NLs de gats sans i comparar les mesures obtingudes amb les tècniques d’imatge amb els valors anatòmics; (ii) caracteritzar els NLs en gats malalts utilitzant TC i US; i (iii) avaluar la capacitat de cada tècnica de d'imatge (TC i US) per discriminar entre canvis neoplàsics i inflamatoris en els NLs.

El nombre dels NLs identificats amb TC va ser més gran que amb US i que en l’estudi anatòmic. Les mesures obtingudes amb TC eren en general més altes que les obtingudes en l'anatomia i US. A més, les comparacions van mostrar diferències significatives entre les tècniques en les mesures, especialment l'altura. L’ús de reconstrucció multiplanar en TC va evitar la sobreestimació de la longitud i l’altura detectada en els talls transversals. Els NLs identificats amb TC van mostrar una aparença similar a la descrita en el gos, essent freqüentment isoatenuats o lleugerament hipoatenuats respecte a la musculatura que els envolta i amb realç homogeni de contrast.

En US, la majoria dels NLs van ser isoecoics o hipoecoics respecte al greix circumdant i amb forma allargada o arrodonida. Alguns NLs es van observar amb perifèria hipoecoica/isoecoica i centre hiperecoic en comparació amb la musculatura, a causa de la presència de greix a l’hili.

La longitud, l'alçada i la ratio S/L, obtinguts amb TC i US, a més de les unitats Hounsfield (HU) abans de l'administració de contrast, van mostrar diferències estadísticament significatives entre els grups control i de casos per a les regions del tòrax i extremitat anterior i abdomen extremitat posterior. A la regió del cap i el coll, les diferències estadístiques entre els grups de control i de casos es van observar només per a l'HU abans i

204 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

després de l'administració de contrast i de la longitud i l'altura obtinguda per US. La ratio S/L obtingut per ambdues tècniques no va mostrar diferències entre les categories inflamatòries i neoplàsiques per a la regió de l'abdomen i les extremitats posteriors, però si en la regió del tòrax i les extremitats anteriors únicament amb TC. Els NLs neoplàsics van ser freqüentment hipoatenuats i heterogenis. Mentre que els inflamatoris van presentar distribució d’atenuacions amb freqüències similars al grup control. L'aspecte heterogeni i hipoatenuat freqüent en els NLs neoplàsics i alguns inflamatoris abans i després de l'administració de contrast podria suggerir canvis severs en l'estructura interna i la vascularització del NL, tot i que no va ser possible determinar un grau determinat de malignitat amb TC.

En US, els NLs neoplàsics van ser freqüentment hipoecoics o heterogenis. Els NLs inflamatoris van ser freqüentment hipoecoics. A la regió del cap i el coll, els NLs neoplàsics van ser més heterogenis amb més freqüència i els NLs inflamatoris més freqüentment hipoecoics.

Els NLs neoplàsics van ser freqüentment arrodonits i amb marges irregulars a diferència dels nóduls inflamatoris i del grup control, els quals sovint eren fusiformes. Els NLs en la categoria d’inflamatoris sovint presentaven marges regulars. Tanmateix, es va observar marges irregulars en NLs molt augmentats de mida, especialment en la categoria de neoplàsia.

En conclusió, els NLs del gat és poden identificar amb les dues tècniques. La tomografia computada va ser més precisa en la detecció i realització de mesures dels LNs que l’US. No obstant això, la superposició de les característiques en ambdues tècniques en les categories del grup cas impedeix una distinció precisa entre inflamació i neoplàsia.

9. RESUMEN

RESUMEN 207

Los nódulos linfáticos (NLs) son estructuras importantes en el diagnóstico y pronóstico de enfermedades tanto neoplásicas como infecciosas. Estudios previos han asociado los cambios observados por ecografía respecto al tamaño (aumento del ratio eje-corto-eje-largo(S/L) a >0,5), la forma (de alargada a redondeada) y la estructura interna de los NLs en humanos y perros con linfadenopatía. Sin embargo, existen muy pocos estudios en gatos en los que se describan estos parámetros con tomografía computarizada (TC) o ecografía (US). Los objetivos de esta tesis fueron: (i) evaluar la capacidad de la TC y la US para identificar NLs de gatos sanos y comparar las mediciones de imagen con valores anatómicos obtenidos; (ii) caracterizar los NLs de gatos enfermos utilizando TC y US; y (iii) evaluar la capacidad de cada técnica de imagen (TC y US) para discriminar entre cambios neoplásicos e inflamatorios de los NLs.

El número de NLs identificados fue mayor con TC que con US y anatomía. Las mediciones obtenidas con TC generalmente fueron más altas que las obtenidas en anatomía y US. Además, las comparaciones mostraron diferencias significativas entre las técnicas en las mediciones, especialmente la altura. El uso de la reconstrucción multiplanar en TC evitó la sobreestimación de la longitud y la altura detectada en los cortes transversales. Los LNs identificados con TC mostraron una apariencia similar a la descrita en el perro, siendo frecuentemente isoatenuantes o ligeramente hipoatenuantes respecto a la musculatura circundante y con captación homogénea de contraste. En US, la mayoría de los NLs fueron isoecoicos o hipoecoicos respecto a la grasa y con forma alargada o redondeada. Algunos LNs se observaron con periferia hipo/isoecoica y centro hiperecoico en comparación con la musculatura, debido a la presencia de grasa en el hilio.

La longitud, la altura y la ratio S/L obtenidos con TC y US, además de las unidades Hounsfield (HU) antes de la administración de contraste, fueron estadísticamente significativas entre los grupos control y caso para las regiones del tórax y miembro anterior, y abdomen y miembro posterior. En la

208 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

región de la cabeza y el cuello, las diferencias estadísticas entre los grupos control y caso fueron observadas solamente para las HU antes y después de la administración de contraste y para la longitud y la altura obtenidas por US. La ratio S/L obtenido con ambas técnicas no mostró diferencias entre las categorías inflamatorias y neoplásicas para la región del abdomen y extremidades posteriores, pero si en la región del tórax y extremidades anteriores únicamente con TC. Los NLs neoplásicos fueron frecuentemente hipoatenuantes y heterogéneos. Mientras que, los inflamatorios presentaron distribución de atenuaciones con frecuencias similares al grupo control. La apariencia heterogénea e hipoatenuante en los NLs neoplásicos y algunos inflamatorios antes y después del contraste podría sugerir cambios importantes en la estructura interna y la vascularización del NL, aunque no se pudo identificar el grado determinado de malignidad con TC.

En US, los NLs neoplásicos fueron frecuentemente hipoecoicos o heterogéneos. Los NLs inflamatorios fueron frecuentemente hipoecoicos. Para la región de la cabeza y el cuello, los NLs neoplásicos fueron más frecuentemente heterogéneos y los NLs inflamatorios más frecuentemente hipoecoicos. Los NLs neoplásicos fueron más frecuentemente redondeados que los inflamatorios y los del grupo control, los cuañes eran más frecuentemente fusiformes. Los NLs en la categoría de inflamatorios presentaban con frecuencia márgenes regulares. Sin embargo, se observaros márgenes irregulares en NLs muy aumentados de tamaño, especialmente en la catergoría de neoplasia.

En conclusión, los NLs del gato se pueden identificar con ambas técnicas. La tomografía computarizada fue más precisa en la detección y realización de mediciones de los NLs que la US. Sin embargo, la superposición de las características en ambas técnicas en las categorías del grupo caso no permitió una distinción precisa entre la gatos con inflamación o neoplasia.

10. OTHER STUDIES DERIVED FROM THIS THESIS

Abstract presented as an oral communication at the EVDI annual meeting, Utrecht (the Netherland), August 2014:

Tobón Restrepo, M., Novellas, R., Espada, Y., Aguilar, A., Moll, X., & Boroffka, S. (2015b). Ultrasonography and computed tomography features of lymph nodes in healthy cats. In Abstracts from the 2014 european veterinary diagnostic imaging annual conference. Veterinary Radiology & Ultrasound (Vol. 56, p. 700). Utrecht, The Netherlands.

Abstract presented as poster at the SEVC-AVEPA 2014, Barcelona, Spain. October 2014: Computed tomography characteristics of the Os penis in the cat

COMPUTED TOMOGRAPHY CHARACTERIZATION OF THE OS PENIS IN THE CAT

Tobón Restrepo M, Espada Y, Altazurra R, Domínguez E, Novellas R.

Objectives: To assess the ability of computed tomography (CT) to visualize the os penis (OP) in the cat.

Materials and Methods: Computed tomography from cats that underwent an abdominal or pelvic CT before and following intravenous administration of contrast medium at the Fundació Hospital Clínic Veterinari from Universitat Autònoma de Barcelona, between October 2013 and April 2014 were reviewed retrospectively. Cats were included in the study if (1) they were no affected by any urinary disease and (2) the external genitals were completely included in the scan. Images were reviewed using a soft tissue and bone algorithms. Width and thickness were measured using an electronic calliper in transverse images. Length was determined using two previously reported methods1, (i) with an electric calliper in sagittal images obtained in the multiplanar reconstructions in both algorithms, (ii) and multiplying slice thickness by the number of consecutive transverse slices that contained the OP. Attenuation values were also measured using a round or oval region of interest (ROI) including as much as possible of the OP, without including other tissues. Groups of ages and intact/castrated condition were used as independent variables to compare between cats with and without visible OP. Chi-square, T-test, Pearson correlation and one-way ANOVA with a 95% confidence interval were used in the statistical analysis.

Results: Twelve cats were included in the study. A cylindrical bone-tissue attenuating structure inside the glans penis compatible with the OP was visible in 11/12 (91.7%) cats. Mean age of the animals with visible OP was 6.2 years (range 0.25 – 11). The OP could not be identified in a 13-year-old castrated cat (8.3%). There was no statistical difference for the visualization of an OP between intact (4/12) vs. castrated (8/12) cats (P = 0.460) or between age groups (≤5y: 6/12; 6-10y: 3/12; ≥11y: 3/12) (P = 0.195). All identified OP had a cylindrical shape. Mean OP width and thickness values in soft tissue (1.45mm; 1.44mm) and bone (1.27mm; 1.08mm) algorithms showed no statistical difference and no correlation with groups of age and intact/castrate condition. Mean OP length showed similar values with both estimation methods (i. Soft 3.46mm Bone 2.9mm; ii.3.4mm) and no statistical difference was found. Mean attenuation before contrast was 230 Hounsfield Units (HU) (range 109 – 402 HU) for soft tissue and 292 HU (range 117 – 528 HU) for bone. Mean attenuation after contrast was 293 HU (range 138 – 427 HU) for soft tissue and 392 HU (range 239 – 682 HU) for bone. No correlation between attenuation values in both algorithms with groups of ages and intact/castrate condition was found.

Discussion and Conclusions: A 5 – 8 mm long os penis without a ventral groove has been reported in the normal cat2. In the present study, cats showed a shorter cylindrical bone with no presence of a ventral groove. In a previous study, a statistical difference in the identification of OP when comparing analog vs. digital X-rays, with 16% and 38% frequencies respectively was found3. In this study, CT showed a frequency of identification

of the OP of 91.7%, which is higher than both analog and digital X-rays reports. No statistical difference between groups of ages and intact/castrate condition with the attenuation values may suggest that the OP mineralization starts early and continues even if the animals have been castrated. On the other hand, slightly higher values in the attenuation values of OP in post contrast images might be due to the presence of contrast agent inside the vascularization of corpus cavernosus that contains the bone1, 4. Even though ROI measurements were placed carefully, Corpus cavernosus tissue might be included due to the OP small diameter. Cats in this study have no signs or evidence of urinary tract disease; this is an important clinical fact because OP might be misinterpreted as signs of urethroliths or dystrophic mineralization of the urethral wall3. Lack of confirmation by histopathology of the OP dimensions is a limitation due to the retrospective nature of the study. For authors knowledge this is the first report of the os penis characteristics by CT in the cat.

1. Nemanic S, Nelson NC. Ultrasonography and noncontrast computed tomography of medial retropharyngeal lymph nodes in healthy cats. American Journal or Veterinary Research 2012; 73(9): 1377 – 1385.

2. König E, Liebich HG: Male genital organs. In König-Liebich (ed): Veterinary anatomy of domestic mammals: textbook and colour atlas. 4th edition. Stuttgart, Editorial Schattauer, 2009, 407 - 422.

3. Piola V, Posch B, Aghte P, et al.: Radiographic characterization of the os penis in the cat. Veterinary Radiology & Ultrasound 2011; 52(3): 270 – 272.

4. Dyce KM, Sack WO, Wensing CJG. The pelvis and reproductive organs of the dog and cat. In Dyce KM, Sack WO, Wensing CJG (eds): Textbook of veterinary anatomy. 4th edition. Philadelphia, Saunders, 2010, 454 – 475.

Abstract accepted as poster at the SEVC-AVEPA 2015, Barcelona, Spain. October 2015: Diagnosis of gallbladder agenesis in a cat using ultrasonography, computed tomography and abdominal laparoscopy.

Book Chapter

Tobón Restrepo, M. (2015). Ultrasound of the abdominal cavity, lymph nodes and large vessels. In R. Novellas Torroja, E. Dominguez Miño, Y. Espada Gerlach, Y. Martínez Pereira, & M. Tobón Restrepo (Eds.), Diagnostic ultrasound in cats (First edit., pp. 211 – 229). Zaragoza, S: Servet editorial - Grupo Asís Biomedia S.L.

11. REFERENCES

REFERENCES 249

Agthe, P., Caine, A. R., Posch, B., & Herrtage, M. E. (2009). Ultrasonographic Appearance of Jejunal Lymph Nodes in Dogs Without Clinical Signs of Gastrointestinal Disease. Veterinary Radiology & Ultrasound, 50(2), 195–200.

Beukers, M., Vilaplana Grosso, F., & Voorhout, G. (2013). Computed Tomographic Characteristics of Presumed Normal Canine Abdominal Lymph Nodes. Veterinary Radiology & Ultrasound, 54(6), 610–617.

Bezuidenhout, A. J. (2013). The lymphatic system. In H. E. Evans & A. De Lahunta (Eds.), Miller’s anatomy of the dog (Forth edit., pp. 535 – 562). St. Louis, Mo: Elsevier.

Bushberg, J. T., Seibert, J. a, Leidholdt, E. M., & Boone, J. M. (2002a). Computed tomography. In The essential physics of medical imaging (Second edit., pp. 321–372). Philadelphia: Lippincott Williams & Wilkins.

Bushberg, J. T., Seibert, J. A., Leidholdt, E. M., & Boone, J. M. (2002b). Ultrasound. In The essential physics of medical imaging (Second edit., pp. 469 – 553). Philadelphia, NJ: Lippincott Williams & Wilkins.

D’Anjou, M.-A. (2008). Abdominal cavity, Lymph nodes, and Great Vessels. In D. Penninck & M.-A. D’Anjou (Eds.), Atlas of small animal ultrasonography (First edit., pp. 445 – 463). Ames, I: Blackwell Publishing Ltd.

D’Anjou, M.-A. (2013). Principles of computed tomography and magnetic resonance imaging. In D. E. Thrall (Ed.), Textbook of veterinary diagnostic radiology (Sixth edit., pp. 50 – 73). St. Louis, Mo: Elsevier.

Drost, W. T. (2013). Physics of the ultrasound imaging. In D. E. Thrall (Ed.), Textbook of veterinary diagnostic radiology (Sixth edit., pp. 38 – 49). St. Louis, Mo: Elsevier.

Dyce, K. M., Sack, W. O., & Wensing, C. J. G. (2002). Textbook of veterinary anatomy. Philadelphia, NJ: Saunders.

Gendler, A., Lewis, J. R., Reetz, J. A., & Schwarz, T. (2008). Computed tomographic features of oral squamous cell carcinoma in cats: 18 cases (2002–2008). Journal of the American Veterinary Medical Association 319–325.

Henninger, W. (2003). Use of computed tomography in the diseased feline thorax. Journal of Small Animal Practice, 44(2), 56–64.

Karnik, K., Reichle, J. K., Fischetti, A. J., & Goggin, J. M. (2009). Computed Tomographic Findings of Fungal Rhinitis and Sinusitis in Cats. Veterinary Radiology & Ultrasound, 50(1), 65–68.

Kealy, J. K., McAllister, H., & Graham, J. P. (2011a). Soft Tissues. In J. K. Kealy, H. McAllister, & J. P. Graham (Eds.), Diagnostic radiology and ultrasonography of the dog and cat (Fifth edit., pp. 543 – 562). St. Louis, Mo: Elsevier Saunders.

Kealy, J. K., McAllister, H., & Graham, J. P. (2011b). The radiograph. In J. K. Kealy, H. McAllister, & J. P. Graham (Eds.), Diagnostic radiology and ultrasonography of the dog and cat (Fifth edit., pp. 10–22). St. Louis, Mo: Elsevier Saunders.

Khanna, R., Sharma, A. D., Khanna, S., Kumar, M., & Shukla, R. C. (2011). Usefulness of ultrasonography for the evaluation of cervical lymphadenopathy. World Journal of Surgical Oncology, 9(1), 29.

Kinns, J., & Mai, W. (2007). Association between malignancy and sonographic heterogeneity in canine and feline abdominal lymph nodes. Veterinary Radiology & Ultrasound, 48(6), 565–569.

Lee, N., Won, S., Choi, M. M., Kim, J., Yi, K., Chang, D., Yoon, J. (2012). CT thoracic duct

250 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

lymphography in cats by popliteal lymph node iohexol injection. Veterinary Radiology & Ultrasound, 53(2), 174–180.

Leinonen, M. R., Raekallio, M. R., Vainio, O. M., Ruohoniemi, M. O., Biller, D. S., & O´Brien, R. T. (2010). Quantitative contrast-enhanced ultrasonographic analysis of perfusion in the kidneys, liver, pancreas, small intestine, and mesenteric lymph nodes in healthy cats. American Journal of Veterinary Research, 71(11), 1305–1311.

Llabres-Diaz, F. J. (2004). Ultrasonography of the medial iliac lymph nodes in the dog. Veterinary Radiology & Ultrasound, 45(2), 156–165.

Mattoon, J. S., Berry, C. R., & Nyland, T. G. (2015). Abdominal Ultrasound Scanning Techniques. In J. S. Mattoon & T. G. Nyland (Eds.), Small Animal Diagnostic Ultrasound (Third edit., pp. 94 – 127). St. Louis, Mo: Elsevier.

Mattoon, J. S., & Nyland, T. G. (2015). Fundamentals of diagnostic ultrasound. In J. S. Mattoon & T. G. Nyland (Eds.), Small Animal Diagnostic Ultrasound (Third edit., pp. 1– 49). St. Louis, Mo: Elsevier.

Mohseni, S., Shojaiefard, A., Khorgami, Z., Alinejad, S., Ghorbani, A., & Ghafouri, A. (2014). Peripheral lymphadenopathy: Approach and diagnostic tools. Iranian Journal of Medical Sciences, 39(2 SUPPL.), 158–170.

Nakamura, K., Takagi, S., Sasaki, N., Bandula Kumara, W. R., Murakami, M., Ohta, H., … Takiguchi, M. (2010). Contrast-enhanced ultrasonography for characterization of canine focal liver lesions. Veterinary Radiology and Ultrasound, 51(1), 79–85.

NAV. (2012). Nomina anatomica veterinaria. ( veterinary gross anatomical nomenclature International committee, Ed.) (Fifth Edit.). Oslo: ICVGAN.

Nemanic, S., & Nelson, N. C. (2012). Ultrasonography and noncontrast computed tomography of medial retropharyngeal lymph nodes in healthy cats. American Journal of Veterinary Research, 73(9), 1377–1385.

Nyman, H. T., Kristensen, A. T., Flagstad, A., & Mcevoy, F. J. (2004). A review of the sonographic assessment of tumor metastases in liver and superficial lymph nodes. Veterinary Radiology & Ultrasound, 45(5), 438–448.

Nyman, H. T., Kristensen, A. T., Skovgaard, I. M., & McEvoy, F. J. (2005). Characterization of normal and abnormal canine superficial lymph nodes using gray-scale B-mode, color flow mapping, power, and spectral doppler ultrasonography: A multivariate study. Veterinary Radiology and Ultrasound, 46(5), 404–410.

Nyman, H. T., & O’Brien, R. T. (2007). The Sonographic Evaluation of Lymph Nodes. Clin Tech Small Anim Pract, 22(3), 128–137.

Ohlerth, S., & Scharf, G. (2007). Computed tomography in small animals - Basic principles and state of the art applications. Veterinary Journal, 173(2), 254–271.

Oliveira, C. R., O’Brien, R. T., Matheson, J. S., & Carrera, I. (2012). Computed tomographic features of feline nasopharyngeal polyps. Veterinary Radiology & Ultrasound, 53(4), 406–11.

Patsikas, M. N., Papadopoulou, P. L., Charitanti, A., Kazakos, G. M., Soultani, C. B., Tziris, N. E., … Stamoulas, K. G. (2010). Computed Tomography and Radiographic Indirect Lymphography for Visualization of Mammary Lymphatic Vessels and the Sentinel Lymph Node in Normal Cats. Veterinary Radiology & Ultrasound, 51(3), 299–304.

Rossi, F., Leone, V. F., Vignoli, M., Laddaga, E., & Terragni, R. (2008). Use of contrast- enhanced ultrasound for characterization of focal splenic lesions. Veterinary Radiology and Ultrasound, 49(2), 154–164.

REFERENCES 251

Rossi, F., Patsikas, M. N., & Wisner, E. R. (2011). Abdominal lymph nodes and lymphatic collecting system. In T. Schwarz & J. H. Saunders (Eds.), veterinary computed tomography (pp. 371 – 379). Ames : Iowa: Wiley-Blackwell.

Saar, L. I., & Getty, R. (1982). Sistema linfático de los carnívoros. In R. Getty (Ed.), S.Sisson - J.D.Grossman. Anatomía de los animales domesticos (Fifth edit., pp. 1811–1831). Barcelona : Spain : Masson.

Salwei, R. M., O’Brien, R. T., & Matheson, J. S. (2005). Characterization of lymphomatous lymph nodes in dogs using contrast harmonic and power doppler ultrasound. Veterinary Radiology and Ultrasound, 46(5), 411–416.

Saunders, J. H., & Ohlerth, S. (2011). CT physics and instrumentation - mechanical desing. In T. Schwarz & J. H. Saunders (Eds.), veterinary computed tomography (pp. 1 – 8). Ames : Iowa: Wiley-Blackwell.

Schreurs, E., Vermote, K., Barberet, V., Daminet, S., Rudorf, H., & Saunders, J. H. (2008). Ultrasonographic Anatomy of Abdominal Lymph Nodes in the Normal Cat. Veterinary Radiology & Ultrasound, 49(1), 68–72.

Schwarz, T., & O’Brien, R. T. (2011). CT Acquisition principles. In T. Schwarz & J. H. Saunders (Eds.), veterinary computed tomography (First edit., pp. 9 – 27). Ames, I: Wiley-Blackwell.

Schwarz, T., & Saunders, J. (2011). Veterinary computed tomography. Ames, I: Wiley- Blackwell.

Steinkamp, H. J., Cornehl, M., Hosten, N., Pegios, W., Vogl, T., & Felix, R. (1995). Cervical lymphadenopathy: Ratio of long- to short-axis diameter as a predictor of malignancy. British Journal of Radiology, 68(807), 266–270.

Steinkamp, H. J., Hosten, N., Richter, C., Schedel, H., & Felix, R. (1994). Enlarged cervical lymph nodes at helical CT. Radiology, 191, 795–798.

Sugimura, M., Kudo, N., & Takahata, K. (1955). Studies on the lympnodi of cats: I. Macroscopical observations on the lymphonodi of heads and necks. Japanese Journal of Veterinary Research, 3(2), 90 – 104.

Sugimura, M., Kudo, N., & Takahata, K. (1956). Studies of lymphonodi of cats: II. Macroscopical observations on the lymphonodi of the body surfaces, thoracic and pelvic limbs. Japanese Journal of Veterinary Research, 4(3), 101 – 112.

Sugimura, M., Kudo, N., & Takahata, K. (1958). Studies on the lymphonodi of cats: III. Macroscopical observations on the lymphonodi in the abdominal and pelvic cavities. Japanese Journal of Veterinary Research, 6(2), 69 – 88.

Sugimura, M., Kudo, N., & Takahata, K. (1959). Studies on the lymphonodi of cats: IV. Macroscopical observations on the lymphonodi in the thoracic cavity and supplemental observations on those in the head and neck. Japanese Journal of Veterinary Research, 7(1-4), 27 – 51.

Tobón Restrepo, M. (2015). Ultrasound of the abdominal cavity, lymph nodes and large vessels. In R. Novellas Torroja, E. Dominguez Miño, Y. Espada Gerlach, Y. Martínez Pereira, & M. Tobón Restrepo (Eds.), Diagnostic ultrasound in cats (First edit., pp. 211 – 229). Zaragoza, S: Servet editorial - Grupo Asís Biomedia S.L.

Tohnosu, N., Onoda, S., & Isono, K. (1989). Ultrasonographic evaluation of cervical lymph node metastases in esophageal cancer with special reference to the relationship between the short to long axis ratio (S/L) and the cancer content. JCU J Clin Ultrasound, 17(2), 101–106.

252 Anatomic and pathologic assessment of feline lymph nodes using computed tomography and ultrasonography

Tompkins, M. B. (1993). Lymphoid system. In Atlas of feline anatomy for veterinarians (pp. 113 – 126). Philadelphia [etc.] : W.B. Saunders Company.

Vassallo, P., Wernecke, K., Roos, N., & Peters, P. (1992). Differentiation of benign from malignant superficial lymphadenopathy: the rol of high-resolution US. Radiology, (183), 215–220.

Widmer, W. R., Mattoon, J. S., & Nyland, T. G. (2015). Peritoneal Fluid, Lymph Nodes, Masses, Peritoneal Cavity, Great Vessels Thrombosis and Focused Examinations. In J. S. Mattoon & T. G. Nyland (Eds.), Small Animal Diagnostic Ultrasound (Third edit., pp. 501 – 516). St. Louis, Mo: Elsevier Saunders.

Wunderbaldinger, P. (2006). Problems and prospects of modern lymph node imaging. Modern Lymphnode Imaging, 58(3), 325–337.